Abstract - Home | Nanyang Technological University (DR-NTU)
A longitudinal study on identification of wandering patterns and their trends in elders with dementia residing in Assisted Living FacilitiesAshish KumarInterdisciplinary Graduate SchoolJoint NTU-UBC Research Centre of Excellence in Active Living for the Elderly (LILY)2017A longitudinal study on identification of wandering patterns and their trends in elders with dementia residing in Assisted Living FacilitiesAshish KumarInterdisciplinary Graduate SchoolJoint NTU-UBC Research Centre of Excellence in Active Living for the Elderly (LILY)A thesis submitted to the Nanyang Technological University in partial fulfillment of the requirement for the degree ofDoctor of Philosophy2017AbstractDementia is a chronic disease which affects the brain and its ability to function normally. It leads to the decline of brain functionalities which hampers its abilities in making judgments, processing language, planning, and sometimes leads to abnormal behavior. Wandering is a significant behavioral problem in Person with Dementia (PWD) and is a heavy burden on caregivers in residential and nursing homes. There have been various cases reported of elopement, getting lost and some serious accident arising out of abnormal wandering behavior. Wandering can be manifested in various forms. Martino-Saltzman in 1991 classified independent travel into four patterns, namely direct, random, pacing and lapping. Out of these four patterns, random, pacing and lapping are considered as wandering. The study of wandering in dementia takes this as reference topology for classification. The objective of this thesis is to create an end-to-end framework for wandering management. The author aims to develop set of algorithms which can be used for the indoor as well as the outdoor environment. The author validated the algorithm which resulted in an overall accuracy of 90%, which is more than any other previous methods used for such hybrid condition.The author also studied the qualitative as well as quantitative aspects of wandering. Qualitative analysis of travel patterns and its variability over a longer period is indicative of the cognitive status of PWD. These measures can be used to alert about the deteriorating condition of patients and can be used to ascertain the usefulness of certain treatment. This system can also be used in a rehabilitation center to monitor progress of the patient. AcknowledgementsI would like to take this opportunity to express my sincere gratitude towards my supervisor Prof Lau Chiew Tong, for extending me all the support and freedom to carry out my research. I thank him for all the valuable suggestions which have shaped my research in the most efficient way. It has been very enlightening and enjoyable experience under his supervision. He has also been very patient hearing my concerns and rectifying it in a very humble way. I would like to thank my co-supervisor Prof Maode Ma, for guiding me and giving me a much valuable suggestion on my research. My research would not have been completed without the help and support of my mentor Prof Chan Syin, her valuable suggestions are the basic pillar of my research. I also express my sincere gratitude to Prof William D. Kearns from the University of South Florida for providing me with all the resources to carry on my research. His continuous encouragement and support, his patience and kindness, and his understanding to bring the best out of me have always been a motivation to excel in what I do. I am truly beholden to have an adviser like him. Without his consistent support and help, the research would not have been so enriching and fulfilling. I would like to thank my parents for giving me mental support through all the successes and failures. It was their blessings which always gave me the courage to face all the challenges and made my path easier.Finally, I would like to thank my friends, and colleagues, Abhijit, Saurabh, Sheetal, Manisha, Antareep, Arun, Garima, Vishram and many more friends from IGS student club for adding life to the graduation years. PhD life would not have been as enjoyable without the company of these wonderful people. Table of Contents TOC \o "1-4" \h \z \u Abstract PAGEREF _Toc522628502 \h iAcknowledgements PAGEREF _Toc522628503 \h iiiTable of Contents PAGEREF _Toc522628504 \h vTable Captions PAGEREF _Toc522628505 \h xiFigure Captions PAGEREF _Toc522628506 \h xiiiChapter 1Introduction PAGEREF _Toc522628507 \h 11.1.Background PAGEREF _Toc522628508 \h 21.2.The Problem PAGEREF _Toc522628509 \h 31.3.Significance of Study PAGEREF _Toc522628510 \h 51.4.Purpose of Study PAGEREF _Toc522628511 \h 61.5.Outline PAGEREF _Toc522628512 \h 6Chapter 2Review of Literature PAGEREF _Toc522628513 \h 72.1.What is Wandering PAGEREF _Toc522628514 \h 82.1.1.A scientific definition of wandering PAGEREF _Toc522628515 \h 82.1.2.A Clinical definition of wandering PAGEREF _Toc522628516 \h 82.1.3.A policy-oriented definition of wandering PAGEREF _Toc522628517 \h 92.2.Measures of Wandering PAGEREF _Toc522628518 \h 92.2.1.Observational Approach PAGEREF _Toc522628519 \h 92.2.2.Technological Approach PAGEREF _Toc522628520 \h 112.3.Need and opportunity to measure Wandering PAGEREF _Toc522628521 \h 112.4.Cognitive and Wandering PAGEREF _Toc522628522 \h 122.4.1.The Mini Mental State Exam (MMSE) PAGEREF _Toc522628523 \h 122.4.2.The Revised Algase Wandering Scale (RAWS-CV) PAGEREF _Toc522628524 \h 132.5.Technologies to manage wandering – A survey PAGEREF _Toc522628525 \h 142.5.1.Sensors Modalities PAGEREF _Toc522628526 \h 152.5.2.Environment PAGEREF _Toc522628527 \h 152.5.3.Support for Independent Living PAGEREF _Toc522628528 \h 162.5.4.Activity recognition approach PAGEREF _Toc522628529 \h 162.5.5.Invasiveness PAGEREF _Toc522628530 \h 172.5.6.Constrain/Restrictiveness PAGEREF _Toc522628531 \h 172.5.7.Support for notification System PAGEREF _Toc522628532 \h 172.5.8.Fall detection PAGEREF _Toc522628533 \h 172.6.A study of few important devices PAGEREF _Toc522628534 \h 222.6.1.Activity recognition approach for wandering detection PAGEREF _Toc522628535 \h 222.6.2.Wandering detection in indoor as well as outdoor environment PAGEREF _Toc522628536 \h 222.6.3.A different approach for wandering detection PAGEREF _Toc522628537 \h 232.6.4.Wandering typology identification PAGEREF _Toc522628538 \h 242.6.5.Off the shelves devices for wandering management PAGEREF _Toc522628539 \h 242.7.Adoption of technology by elderly and healthcare professional PAGEREF _Toc522628540 \h 242.8.Motivation PAGEREF _Toc522628541 \h 252.9.Conclusion PAGEREF _Toc522628542 \h 27Chapter 3Classification of Wandering Patterns in Person with Dementia PAGEREF _Toc522628543 \h 293.1.Introduction PAGEREF _Toc522628544 \h 303.2.Related Work PAGEREF _Toc522628545 \h 333.3.Preliminaries PAGEREF _Toc522628546 \h 333.3.1.Physical layout Representation PAGEREF _Toc522628547 \h 333.3.2.Wandering typology in a grid map representation PAGEREF _Toc522628548 \h 343.4.The Proposed Algorithm PAGEREF _Toc522628549 \h 373.4.1.Data Pre-processing PAGEREF _Toc522628550 \h 373.4.1.1.Data Cleaning: PAGEREF _Toc522628551 \h 373.4.1.2.Trajectory Compression: PAGEREF _Toc522628552 \h 383.4.2.Segmentation PAGEREF _Toc522628553 \h 403.4.2.1.Stay-point detection: PAGEREF _Toc522628554 \h 423.4.2.2.Segmentation: PAGEREF _Toc522628555 \h 423.4.3.Classification PAGEREF _Toc522628556 \h 433.4.3.1.Features Calculation: PAGEREF _Toc522628557 \h 433.4.3.2.Episode Classification: PAGEREF _Toc522628558 \h 433.5.A sample episode representation in grid world PAGEREF _Toc522628559 \h 453.5.1.Grid Representation of Layout: PAGEREF _Toc522628560 \h 453.6.Framework (RT-WMAT): Real-time Wandering Management & Analytics Tool PAGEREF _Toc522628561 \h 483.6.1.Goal and Scope of the Framework PAGEREF _Toc522628562 \h 503.6.2.System Architecture PAGEREF _Toc522628563 \h 503.6.2.1.Sensors (Spatiotemporal data and contextual information): PAGEREF _Toc522628564 \h 513.6.2.2.Server (Data storage and Analytics engine): PAGEREF _Toc522628565 \h 513.6.2.3.Clients: PAGEREF _Toc522628566 \h 523.6.3.Implementation and Analysis of results PAGEREF _Toc522628567 \h 533.6.4.Discussions PAGEREF _Toc522628568 \h 543.7.Conclusion and Future Work PAGEREF _Toc522628569 \h 55Chapter 4A Data Analytic Study of Wandering PAGEREF _Toc522628570 \h 574.1.Introduction PAGEREF _Toc522628571 \h 584.2.Methods PAGEREF _Toc522628572 \h 584.2.1.Subjects PAGEREF _Toc522628573 \h 584.2.2.Apparatus PAGEREF _Toc522628574 \h 594.2.3.Description of the monitored area PAGEREF _Toc522628575 \h 614.2.4.Procedure for data collection PAGEREF _Toc522628576 \h 624.2.4.1.Indoor navigation data: PAGEREF _Toc522628577 \h 624.2.4.2.Outdoor navigation data: PAGEREF _Toc522628578 \h 634.3.Wandering Pattern Identification PAGEREF _Toc522628579 \h 664.3.1.Procedure PAGEREF _Toc522628580 \h 664.3.2.Verification of algorithm using controlled experiment PAGEREF _Toc522628581 \h 664.3.3.Results PAGEREF _Toc522628582 \h 674.3.4.Discussion PAGEREF _Toc522628583 \h 684.4.Statistical analysis of the Wandering Pattern PAGEREF _Toc522628584 \h 694.4.1.Procedure PAGEREF _Toc522628585 \h 694.4.2.Results PAGEREF _Toc522628586 \h 694.4.2.1.Relationship of MMSE Spatial and Temporal Orientation to the wandering pattern: PAGEREF _Toc522628587 \h 724.4.2.2.Logistic regression to identify dementia diagnostic group: PAGEREF _Toc522628588 \h 724.4.2.3.Contrast of the proposed algorithm vs. Fractal Dimension: PAGEREF _Toc522628589 \h 734.4.3.Discussion: PAGEREF _Toc522628590 \h 744.5.Conclusion PAGEREF _Toc522628591 \h 75Chapter 5Analysis of trend in the navigation PAGEREF _Toc522628592 \h 775.1.Introduction PAGEREF _Toc522628593 \h 785.2.Motivation PAGEREF _Toc522628594 \h 785.3.Methods PAGEREF _Toc522628595 \h 795.3.1.The Proposed Algorithm PAGEREF _Toc522628596 \h 795.3.2.Features Generation PAGEREF _Toc522628597 \h 805.3.3.Trend Analysis in Navigation PAGEREF _Toc522628598 \h 825.4.Subject and Layout PAGEREF _Toc522628599 \h 855.5.Results and Discussion PAGEREF _Toc522628600 \h 885.5.1.Trend Analysis PAGEREF _Toc522628601 \h 885.5.2.Distinction between two diagnostic groups: PAGEREF _Toc522628602 \h 895.6.Conclusion PAGEREF _Toc522628603 \h 89Chapter 6Summary and Future work PAGEREF _Toc522628604 \h 906.1.Revisiting the research objectives PAGEREF _Toc522628605 \h 926.2.Limitation of our research PAGEREF _Toc522628606 \h 936.3.Adoption of technology for wandering management PAGEREF _Toc522628607 \h 946.4.Future Work PAGEREF _Toc522628608 \h 956.4.1.Collection of larger dataset for longitudinal study PAGEREF _Toc522628609 \h 956.4.2.Activity recognition approach for contextual wandering analysis PAGEREF _Toc522628610 \h 956.4.3.Personalized User wandering detection PAGEREF _Toc522628611 \h 95Appendix PAGEREF _Toc522628612 \h 97List of publication PAGEREF _Toc522628613 \h 97Under Review PAGEREF _Toc522628614 \h 97References PAGEREF _Toc522628615 \h 98Table Captions TOC \h \z \c "Table" Table 2.1: MMSE and Cognition PAGEREF _Toc522628347 \h 13Table 2.2: Scale used in RAWS-CV PAGEREF _Toc522628348 \h 13Table 2.3: Sub-scales of RAWS-CV PAGEREF _Toc522628349 \h 14Table 2.4: Event monitoring-based Wandering detection PAGEREF _Toc522628350 \h 19Table 2.5: Trajectory tracking-based wandering detection PAGEREF _Toc522628351 \h 20Table 2.6: Location-based prevention of wandering-related adverse events PAGEREF _Toc522628352 \h 21Table 3.1: Features of the episode PAGEREF _Toc522628353 \h 43Table 3.2: Coding of the navigation direction PAGEREF _Toc522628354 \h 47Table 3.3: Percentage of navigational pattern during time of day (ToD) PAGEREF _Toc522628355 \h 53Table 4.1: Classification Table for the controlled experiment PAGEREF _Toc522628356 \h 67Table 4.2: Classification Table PAGEREF _Toc522628357 \h 68Table 4.3: Fractional distribution of navigational pattern in subjects staying in ALF PAGEREF _Toc522628358 \h 71Table 4.4: Classification Table PAGEREF _Toc522628359 \h 73Table 4.5: Correlations of Fractal D and MMSE with Spatial Measures of Wandering PAGEREF _Toc522628360 \h 74Table 4.6: Correlations of Fractal D and MMSE with Temporal Measures of Wandering PAGEREF _Toc522628361 \h 74Table 5.1: Grouping of features in chronological order PAGEREF _Toc522628362 \h 84Table 5.2: Subject demographic PAGEREF _Toc522628363 \h 86Table 5.3: Distribution of wandering pattern PAGEREF _Toc522628364 \h 87Figure Captions TOC \h \z \c "Figure" Figure 1.1: Percentage change in the world’s population by age: 2010-2050* PAGEREF _Toc522628391 \h 2Figure 2.1: Parameters used to evaluate the wandering technology PAGEREF _Toc522628392 \h 15Figure 3.1: Current field of research in Wandering PAGEREF _Toc522628393 \h 30Figure 3.2: Patterns of ambulation (Martino-Salzman, 1991) PAGEREF _Toc522628394 \h 31Figure 3.3: Direction of navigation from grid PAGEREF _Toc522628395 \h 36Figure 3.4: Navigational Pattern in grid world representation PAGEREF _Toc522628396 \h 36Figure 3.5: The Workflow Framework PAGEREF _Toc522628397 \h 37Figure 3.6: Example of trajectory compression PAGEREF _Toc522628398 \h 40Figure 3.7: Step Discretization method in episode plotting PAGEREF _Toc522628399 \h 42Figure 3.8: Flowchart for episode consolidation PAGEREF _Toc522628400 \h 45Figure 3.9: Layout divided into equally sized grid PAGEREF _Toc522628401 \h 46Figure 3.10: Sample episode plot in Matlab PAGEREF _Toc522628402 \h 46Figure 3.11: A Real-Time Wandering Management and Analytics Tool (RT-WMAT) PAGEREF _Toc522628403 \h 51Figure 3.12: Trend in navigation pattern PAGEREF _Toc522628404 \h 54Figure 3.13: Trend diagram over a year PAGEREF _Toc522628405 \h 55Figure 4.1: A Ubisense Compact Tag PAGEREF _Toc522628406 \h 59Figure 4.2: Full Specification of Ubisense Series 7000 Compact Tag PAGEREF _Toc522628407 \h 60Figure 4.3: Floor plans for research site 1 (top) and 2 (bottom); sensor locations are at the vertices of the shaded regions and the origin is in the lower left; major divisions are 10m increments; individual participants appear as numbered ovals PAGEREF _Toc522628408 \h 61Figure 4.4: Data collection method for outdoor navigation using GPS sensor from mobile PAGEREF _Toc522628409 \h 63Figure 4.5: Data structure used for episode classification PAGEREF _Toc522628410 \h 65Figure 4.6: Precision and recall of the proposed algorithm PAGEREF _Toc522628411 \h 68Figure 5.1: The Complete Framework PAGEREF _Toc522628412 \h 80Figure 5.2: A sample trajectory path PAGEREF _Toc522628413 \h 81Figure 5.3: Trend in features over a 1-year duration PAGEREF _Toc522628414 \h 88Figure 5.4: Box plot for Dementia and Control Group PAGEREF _Toc522628415 \h 89Figure 6.1: Phases in wandering PAGEREF _Toc522628416 \h 95IntroductionThis chapter introduces the background and purpose behind the study into the wandering science. Subsequently, the author also presents the significance and objectives of the research. The author discusses the problem faced by the ageing population and how this research is relevant to them. By the end of the chapter, the research problems are formulated which are addressed in the subsequent chapters.BackgroundWith improvements in public health, nutrition and medical facilities, there has been a dramatic increase in average life expectancy of people. With advanced medical facilities available round the clock in Singapore and other developed countries, longevity and life expectancy of citizens have increased considerably and are expected to increase even further. On one side, it is good for the society, but on the other side it faces some dire challenges associated with old age. According to the world population survey the proportion of old and oldest-old (people aged 85 or higher) is expected to rise at a much higher rate as shown in REF _Ref504558165 \h Figure 1.1.Figure 1.1: Percentage change in the world’s population by age: 2010-2050*Old people often have limited regenerative abilities and are more susceptible to disease, syndromes, and sickness than younger adults. Old age comes with a plethora of ailments and disabilities, most of them being chronic requiring continuous care and financial stability. A primary goal of sustainable development is for people to live longer, more productive, more satisfying lives. Independent living and assisted living facilities *Source: United Nations, World Population Prospects: The 2010 Revision.Available at: the means to take care of elderly people. However, these facilities are burdened with an enormous explosion of population and are very costly to be affordable for the general population. Dementia is one of the most prevalent diseases associated with old age. It is generally manifested by loss of cognitive function, memory, reasoning, and speech impairment. It is very hard to detect dementia at an early stage because most of the problem associated with the disease is attributed to the normal effects of aging. This places a growing demand for health and long-term care provider for the development of prevention technologies. The cause of dementia is largely unknown, but in the final stage there is considerable loss of cognitive ability. Wandering is generally associated with the decline of cognitive abilities in person. Thus, wandering is associated with one of the first symptoms of dementia. Our study is focused on the identification of the wandering episode in person. It can be identified by analyzing the navigation pattern of a person. By analyzing these patterns over a longer period, it can be used to predict the onslaught of dementia at an early stage. The Problem Increasing life expectancy resulting from better medical services and health awareness, coupled with a falling birth-rate have led to a significant rise in the number of elderly citizens in developed nations. The proportion of elderly in the worldwide population is expected to increase from 8% in 2010 to 16% by 2050. Increasingly aged populations and fewer caregivers have required many countries to embrace technology to provide better medical and long-term care support for the elderly, while at the same time employing fewer workers. Elderly persons are prone to various ailments leading to functional declines in physical strength and cognition, expressed as poor motor coordination, short and/or long-term memory declines, and identity confusion and/or impaired judgment ADDIN EN.CITE <EndNote><Cite><Author>Wentzel</Author><Year>2001</Year><RecNum>174</RecNum><DisplayText>[1]</DisplayText><record><rec-number>174</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">174</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Wentzel, C</author><author>Rockwood, K</author><author>MacKnight, C</author><author>Hachinski, V</author><author>Hogan, DB</author><author>Feldman, H</author><author>?stbye, T</author><author>Wolfson, C</author><author>Gauthier, S</author><author>Verreault, R</author></authors></contributors><titles><title>Progression of impairment in patients with vascular cognitive impairment without dementia</title><secondary-title>Neurology</secondary-title></titles><periodical><full-title>Neurology</full-title></periodical><pages>714-716</pages><volume>57</volume><number>4</number><dates><year>2001</year></dates><isbn>0028-3878</isbn><urls></urls></record></Cite></EndNote>[1]. It is estimated that 17% elderly are cognitively impaired due to diseases such as Alzheimer’s disease, Parkinson’s dementia, or dementia with Lewy bodies ADDIN EN.CITE <EndNote><Cite><Author>Rodrigues</Author><Year>2012</Year><RecNum>154</RecNum><DisplayText>[2]</DisplayText><record><rec-number>154</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">154</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Rodrigues, Ricardo</author><author>Huber, Manfred</author><author>Lamura, Giovanni</author></authors></contributors><titles><title>Facts and figures on healthy ageing and long-term care</title><secondary-title>Vienna: European Centre for Social Welfare Policy and Research</secondary-title></titles><periodical><full-title>Vienna: European Centre for Social Welfare Policy and Research</full-title></periodical><dates><year>2012</year></dates><urls></urls></record></Cite></EndNote>[2], and this number is expected to climb to 21% by the year 2030 ADDIN EN.CITE <EndNote><Cite><Author>Suzman</Author><Year>2015</Year><RecNum>169</RecNum><DisplayText>[3]</DisplayText><record><rec-number>169</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">169</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Suzman, Richard</author><author>Beard, John R</author><author>Boerma, Ties</author><author>Chatterji, Somnath</author></authors></contributors><titles><title>Health in an ageing world—what do we know?</title><secondary-title>The Lancet</secondary-title></titles><periodical><full-title>The Lancet</full-title></periodical><pages>484-486</pages><volume>385</volume><number>9967</number><dates><year>2015</year></dates><isbn>0140-6736</isbn><urls></urls></record></Cite></EndNote>[3]. The US Centres for Disease Control and Prevention rank Alzheimer's disease as the fifth leading cause of death in those aged 65 – 85 years, resulting in 92,604 deaths in 2014 alone. Wandering is a serious behavioral symptom characteristic of dementias, particularly Alzheimer's disease and vascular dementia; and reports indicate 67% of persons with dementia may exhibit wandering behavior as the disease progresses ADDIN EN.CITE <EndNote><Cite><Author>Silverstein</Author><Year>2002</Year><RecNum>158</RecNum><DisplayText>[4]</DisplayText><record><rec-number>158</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">158</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Silverstein, NM</author><author>Flaherty, G</author><author>Tobin, TS</author></authors></contributors><titles><title>Comprehensive review of the literature on wandering behavior</title><secondary-title>Dementia and wandering behavior: concern for the lost elder. New York: Springer</secondary-title></titles><periodical><full-title>Dementia and wandering behavior: concern for the lost elder. New York: Springer</full-title></periodical><pages>28-51</pages><dates><year>2002</year></dates><urls></urls></record></Cite></EndNote>[4]. Wanderers may exhibit physical aggression, restlessness, agitation, and temporal and spatial disorientation ADDIN EN.CITE <EndNote><Cite><Author>Pollock</Author><Year>2007</Year><RecNum>175</RecNum><DisplayText>[5]</DisplayText><record><rec-number>175</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">175</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Pollock, Bruce G</author><author>Mulsant, Benoit H</author><author>Rosen, Jules</author><author>Mazumdar, Sati</author><author>Blakesley, Richard E</author><author>Houck, Patricia R</author><author>Huber, Kimberly A</author></authors></contributors><titles><title>A double-blind comparison of citalopram and risperidone for the treatment of behavioral and psychotic symptoms associated with dementia</title><secondary-title>The American Journal of Geriatric Psychiatry</secondary-title></titles><periodical><full-title>The American Journal of Geriatric Psychiatry</full-title></periodical><pages>942-952</pages><volume>15</volume><number>11</number><dates><year>2007</year></dates><isbn>1064-7481</isbn><urls></urls></record></Cite></EndNote>[5]. Wandering as a behavioral symptom is considered an important indicator of the cognitive state of the person with dementia ADDIN EN.CITE <EndNote><Cite><Author>Hope</Author><Year>2001</Year><RecNum>163</RecNum><DisplayText>[6]</DisplayText><record><rec-number>163</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">163</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Hope, Tony</author><author>Keene, Janet</author><author>McShane, Rupert H</author><author>Fairburn, Christopher G</author><author>Gedling, Kathy</author><author>Jacoby, Robin</author></authors></contributors><titles><title>Wandering in dementia: a longitudinal study</title><secondary-title>International Psychogeriatrics</secondary-title></titles><periodical><full-title>International Psychogeriatrics</full-title></periodical><pages>137-147</pages><volume>13</volume><number>02</number><dates><year>2001</year></dates><isbn>1741-203X</isbn><urls></urls></record></Cite></EndNote>[6]. Martino-Saltzman ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>159</RecNum><DisplayText>[7]</DisplayText><record><rec-number>159</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">159</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7] performed some of the earliest studies of patients’ wandering by categorizing the geometry of their navigational patterns, and this method remains the prevailing strategy for evaluating wandering in dementia. In Martino-Saltzman’s typology, navigation can be categorized into one of four patterns: direct, random, lapping or pacing; the latter three patterns characterizing wandering. Using this typology the episodic frequency and duration of the random pattern, has been found to be moderately correlated with spatial disorientation and cognitive impairment ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2001</Year><RecNum>165</RecNum><DisplayText>[8]</DisplayText><record><rec-number>165</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">165</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author><author>Beattie, Elizabeth RA</author><author>Therrien, Barbara</author></authors></contributors><titles><title>Impact of cognitive impairment on wandering behavior</title><secondary-title>Western Journal of Nursing Research</secondary-title></titles><periodical><full-title>Western Journal of Nursing Research</full-title></periodical><pages>283-295</pages><volume>23</volume><number>3</number><dates><year>2001</year></dates><isbn>0193-9459</isbn><urls></urls></record></Cite></EndNote>[8]. Lapping and pacing are rare events compared to direct and random motion, and research into their significance is scant. In studies of lower animals, lapping may be a strategy for relocating a space, suggesting loss of the spatial memory trace and residual familiarity of space and context, whereas it serves as a wayfinding strategy in mice ADDIN EN.CITE <EndNote><Cite><Author>Holden</Author><Year>1993</Year><RecNum>164</RecNum><DisplayText>[9]</DisplayText><record><rec-number>164</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">164</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Holden, Janean Erickson</author><author>Therrien, Barbara</author></authors></contributors><titles><title>Cue familiarity reduces spatial disorientation following hippocampal damage</title><secondary-title>Nursing research</secondary-title></titles><periodical><full-title>Nursing research</full-title></periodical><pages>338-343</pages><volume>42</volume><number>6</number><dates><year>1993</year></dates><isbn>0029-6562</isbn><urls></urls></record></Cite></EndNote>[9], substantiating human studies that have found only weak associations with levels of cognitive impairment ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2001</Year><RecNum>165</RecNum><DisplayText>[8]</DisplayText><record><rec-number>165</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">165</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author><author>Beattie, Elizabeth RA</author><author>Therrien, Barbara</author></authors></contributors><titles><title>Impact of cognitive impairment on wandering behavior</title><secondary-title>Western Journal of Nursing Research</secondary-title></titles><periodical><full-title>Western Journal of Nursing Research</full-title></periodical><pages>283-295</pages><volume>23</volume><number>3</number><dates><year>2001</year></dates><isbn>0193-9459</isbn><urls></urls></record></Cite></EndNote>[8]. In contrast, pacing has been considered by Neistein and Siegal as representative of agitation and, like lapping, is repetitive in nature ADDIN EN.CITE <EndNote><Cite><Author>Neistein</Author><Year>1997</Year><RecNum>181</RecNum><DisplayText>[10]</DisplayText><record><rec-number>181</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">181</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Neistein, Susan</author><author>Siegal, Alan P</author></authors></contributors><titles><title>Agitation, wandering, pacing, restlessness, and repetitive mannerisms</title><secondary-title>International Psychogeriatrics</secondary-title></titles><periodical><full-title>International Psychogeriatrics</full-title></periodical><pages>399-402</pages><volume>8</volume><number>S3</number><dates><year>1997</year></dates><isbn>1741-203X</isbn><urls></urls></record></Cite></EndNote>[10]. Pacing may serve a twofold purpose: discharging excess energy while providing an emotionally self-soothing mechanism.Early research establishing wandering typologies was of necessity observational in nature; patients’ actions were manually coded and in later studies behaviors were videotaped. The dominant wandering pattern present was captured using predetermined coding sequences ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>159</RecNum><DisplayText>[7]</DisplayText><record><rec-number>159</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">159</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7]. This required considerable time and effort by highly trained observers and constituted an extended vigilance task prone to human error so typically only a very few subjects could be observed simultaneously. Algase found her observers were capable of reliably detecting instances of lapping and pacing (which were repetitive) and direct travel, but were incapable of reliably detecting random travel that, by definition, lacked periodicity. The protocol was reactive and cumbersome, and the constant surveillance impinged on patient privacy. Parameters such as speed and directional changes, and path length were almost impossible to estimate reliably using observational methods.Significance of StudyAttempts to advance the science of wandering pattern detection have increased in the recent year, some gaps still remain. While most of the study focused on the identification of a random pattern, which is only one of the three wandering patterns (random, lapping and pacing), little effort can be found in the scientific community to reliably identify lapping and pacing. Though lapping and pacing form very small fraction of navigation, the study can open the new perspective in the overall wandering management. Studies conducted in the wandering detection have collected a large amount of navigation data, but management of these data for long-term use has never been realized. With the collection of private data in the form of geo-location, the security of such system becomes a major concern. There has been still lack of complete framework which can take care of all the steps from data collection to the analysis of wandering. There has also been lack of unified algorithm which can be applied to all sorts of dataset irrespective of data collected from indoor or outdoor scenario.Due to the lack of wandering management framework, longitudinal study on these data cannot be found in the scientific community. The framework has opened a new gate to analyse these data over a lengthy period to find trends if any in these data. These trends are very important when it comes to the diagnostic of dementia. Few attempts have been made to identify the correlation between features of navigation and various cognitive tests such as MMSE and RAWS-CV. A long-term study in this regard is still missing and attempts to mend this gap is not substantial. A quantitative measure of correlation between wandering behavior per se and its related factors is very crucial for the in-depth analysis of wandering behavior in field research. As wandering behavior occurs for a variety of reasons and is tightly linked to individuals’ health status, wandering history and environmental situations, the measures, criteria, and evaluation of correlation between wandering and the possible related factors become another research challenge. Statistical and data mining methods can be applied to study and extract the hidden correlation and quantitative dependence from large-scale wandering data collected from individuals and group.Purpose of StudyThe purpose of this study is to advance the science of wandering management by providing clinicians and researchers with new instruments that will lead to a fuller comprehension of its nature and origin. Our intention is to automate the data analytics process for wandering which can work across all the sensor modalities. The Author also intends to find any statistically significant trend in the longitudinal data. OutlineThis report is organized as follows: REF _Ref504640614 \r \h Chapter 1 introduces the problem and backgrounds. It further discusses the significance of this study and the research statement. Chapter 2 provides the literature review of some of the related work in this field. Chapter 3 discuss the complete framework and novel grid based wandering pattern detection algorithm.Chapter 4 validates the algorithm and framework on real world data set.Chapter 5 does the longitudinal study on the data to find consistent trend in wandering features.Finally, Chapter 6 summarizes the report with the conclusion and future work. ADDIN EN.SECTION.REFLIST Review of LiteratureThis chapter starts with the definition of wandering. It provides an overview of wandering and its significance in dementia care, as well as the technologies for detecting, measuring, and managing wandering. The author also discusses why measuring wandering is important and what new insight it gives in terms of the cognitive state of a person. In the next section, a systematic literature review of the technologies used in the wandering detection has been performed. Parameters used to evaluate wandering technologies are also presented. Finally, the five main gaps in the existing research has been identified which has been mapped into five research objectives.What is WanderingWandering has been studied for decades encompassing various aspects such as: (1) the nature of wandering defined in terms of its descriptors, measurement, and natural history; (2) the outcomes related with wandering such as getting lost and elopement; (3) wandering related behavior; such as intrusion, shadowing or exit seeking; and (4) techniques for wandering management and intervention. Even after such a long period of study, nature and behaviors remain poorly defined due to the many-faceted nature of wandering. This leads to confusion and researchers taking its own disciplinary and professional viewpoint which results in inconsistencies across various domain. The Precise definition of wandering will work towards improving science, practice, and policy and serve to enhance communication across fields.In this section, wandering is defined specific to three main stakeholders, scientist, clinician, and policy maker.A scientific definition of wanderingIt describes wandering behaviour shaped by four domains: space, time, locomotion, and drive for wandering. This is based on the definition of wandering as defined by Algase ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>1999</Year><RecNum>308</RecNum><DisplayText>[11]</DisplayText><record><rec-number>308</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516188791">308</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author></authors></contributors><titles><title>Wandering in dementia</title><secondary-title>Annual review of nursing research</secondary-title></titles><periodical><full-title>Annual review of nursing research</full-title></periodical><pages>185-217</pages><volume>17</volume><number>1</number><dates><year>1999</year></dates><isbn>0739-6686</isbn><urls></urls></record></Cite></EndNote>[11], which considers wandering behaviour in terms of its volume (frequency, rate, and duration), spatial and temporal distribution, and geographical patterns. It is defined as:“Wandering is a syndrome of dementia-related locomotion behavior having a frequent, repetitive, temporally-disordered, and/or spatially disoriented nature that is manifested in lapping, random, and/or pacing patterns, some of which are associated with eloping, eloping attempts, or getting lost unless accompanied.”This definition of wandering will be of interest as it will be dealing with the measures of wandering in terms of intensity and geographical pattern and it allows others to independently verify the validity of research results through observation, measurement, or testing.A Clinical definition of wanderingThis definition of wandering is of interest for the people who directly interact with dementia patients. This has been defined as “meandering, aimless or repetitive locomotion that exposes a person to harm and is incongruent with boundaries, limits or obstacles. North American Nursing Diagnosis Association (NANDA) lists 13 specific wandering behaviors and conditions ADDIN EN.CITE <EndNote><Cite><Author>Herdman</Author><Year>2011</Year><RecNum>309</RecNum><DisplayText>[12]</DisplayText><record><rec-number>309</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516188944">309</key></foreign-keys><ref-type name="Book">6</ref-type><contributors><authors><author>Herdman, T Heather</author></authors></contributors><titles><title>Nursing diagnoses 2012-14: Definitions and classification</title></titles><dates><year>2011</year></dates><publisher>John Wiley & Sons</publisher><isbn>0470654821</isbn><urls></urls></record></Cite></EndNote>[12] under which wandering occurs. A policy-oriented definition of wanderingIt helps policy makers and risk managers to focus mainly on the safety, risk, and adverse outcome such as getting lost or elopement related with the wandering. It helps in designing the policy for the stakeholders. The Department of Veterans Affairs defines a wandering patient as follows: A wandering patient is a high-risk patient who has shown a propensity to stray beyond the view or control of employees, thereby requiring a high degree of monitoring and protection to ensure the patient’s safety. Measures of WanderingThis section reviews observational, and technological methods to assess, quantify, and measure wandering and wandering-related behaviors. Observational ApproachOne of the most prevalent approaches to measuring wandering is characterized by the geographical pattern of navigation. Martino-Saltzman ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>159</RecNum><DisplayText>[7]</DisplayText><record><rec-number>159</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">159</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7] performed some of the earliest studies of patients’ wandering by categorizing the geometry of their navigational patterns, and this method remains the prevailing strategy for evaluating wandering in dementia. In Martino-Saltzman’s typology, navigation can be categorized into one of four patterns: direct, random, lapping or pacing; These patterns are derived using videotape method from nursing home residents. It is defined as:Direct—locomotion from a point to a destination along a straightforward or uncomplicated path and without significant hesitation.Random—locomotion along a haphazard path with multiple legs and directional changes and hesitations of up to 30 seconds at any point along the path;Lapping—locomotion that has a circuitous path (closed loop) with at least three legs that (a) return the wanderer to his or her point of origin, and (b) may include brief (several seconds) stops or hesitations as the wanderer changes directional heading along the path; and Pacing—back and forth locomotion between two end points, at which directional heading is reversed;Out of these four patterns, only the random, lapping and pacing patterns characterize wandering.In this approach, individual episodes of locomotion are coded for start and stop time using very simple tools such as paper, pencil, and stopwatch. Sometimes, a computer is used for coding information. From coded observations, a variety of metrics is generated for individual travel patterns and for wandering overall. Both frequency and duration of wandering episodes can be derived from observational data, and both are important to consider, as each parameter may yield a somewhat different perspective of wandering. As a result, there have been many insightful results using this approach: This methodology has been extensively used by Algase and colleagues in several descriptive studies of wandering PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BbGdhc2U8L0F1dGhvcj48WWVhcj4yMDAxPC9ZZWFyPjxS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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BbGdhc2U8L0F1dGhvcj48WWVhcj4yMDAxPC9ZZWFyPjxS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ADDIN EN.CITE.DATA [8, 13, 14]. Furthermore, using this typology the episodic frequency and duration of the random pattern has been found to be moderately correlated with spatial disorientation and cognitive impairment ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2001</Year><RecNum>165</RecNum><DisplayText>[8]</DisplayText><record><rec-number>165</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">165</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author><author>Beattie, Elizabeth RA</author><author>Therrien, Barbara</author></authors></contributors><titles><title>Impact of cognitive impairment on wandering behavior</title><secondary-title>Western Journal of Nursing Research</secondary-title></titles><periodical><full-title>Western Journal of Nursing Research</full-title></periodical><pages>283-295</pages><volume>23</volume><number>3</number><dates><year>2001</year></dates><isbn>0193-9459</isbn><urls></urls></record></Cite></EndNote>[8]. Lapping and pacing are rare events compared to direct and random motion, and research into their significance is scant. In studies of lower animals, lapping may be a strategy for relocating a space, suggesting loss of the spatial memory trace and residual familiarity of space and context, whereas it serves as a wayfinding strategy in mice ADDIN EN.CITE <EndNote><Cite><Author>Holden</Author><Year>1993</Year><RecNum>164</RecNum><DisplayText>[9]</DisplayText><record><rec-number>164</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">164</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Holden, Janean Erickson</author><author>Therrien, Barbara</author></authors></contributors><titles><title>Cue familiarity reduces spatial disorientation following hippocampal damage</title><secondary-title>Nursing research</secondary-title></titles><periodical><full-title>Nursing research</full-title></periodical><pages>338-343</pages><volume>42</volume><number>6</number><dates><year>1993</year></dates><isbn>0029-6562</isbn><urls></urls></record></Cite></EndNote>[9], substantiating human studies that have found only weak associations with levels of cognitive impairment ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2001</Year><RecNum>165</RecNum><DisplayText>[8]</DisplayText><record><rec-number>165</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">165</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author><author>Beattie, Elizabeth RA</author><author>Therrien, Barbara</author></authors></contributors><titles><title>Impact of cognitive impairment on wandering behavior</title><secondary-title>Western Journal of Nursing Research</secondary-title></titles><periodical><full-title>Western Journal of Nursing Research</full-title></periodical><pages>283-295</pages><volume>23</volume><number>3</number><dates><year>2001</year></dates><isbn>0193-9459</isbn><urls></urls></record></Cite></EndNote>[8]. In contrast, pacing has been considered by Neistein and Siegal as representative of agitation and, like lapping, is repetitive in nature ADDIN EN.CITE <EndNote><Cite><Author>Neistein</Author><Year>1997</Year><RecNum>181</RecNum><DisplayText>[10]</DisplayText><record><rec-number>181</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">181</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Neistein, Susan</author><author>Siegal, Alan P</author></authors></contributors><titles><title>Agitation, wandering, pacing, restlessness, and repetitive mannerisms</title><secondary-title>International Psychogeriatrics</secondary-title></titles><periodical><full-title>International Psychogeriatrics</full-title></periodical><pages>399-402</pages><volume>8</volume><number>S3</number><dates><year>1997</year></dates><isbn>1741-203X</isbn><urls></urls></record></Cite></EndNote>[10]. Pacing may serve a twofold purpose: discharging excess energy while providing an emotionally self-soothing mechanism.Early research establishing wandering typologies was of necessity observational in nature; patients’ actions were manually coded and in later studies behaviors were videotaped. The dominant wandering pattern present was captured using predetermined coding sequences ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>159</RecNum><DisplayText>[7]</DisplayText><record><rec-number>159</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">159</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7]. This required considerable time and effort by highly trained observers and constituted an extended vigilance task prone to human error so typically only a very few subjects could be observed simultaneously. Algase found her observers were capable of reliably detecting instances of lapping and pacing (which were repetitive) and direct travel, but were incapable of reliably detecting random travel that, by definition, lacked periodicity. The protocol was reactive and cumbersome, and the constant surveillance impinged on patient privacy. Parameters such as speed and directional changes, and path length were almost impossible to estimate reliably using observational methods. Technological ApproachThe study conducted using observational approaches yielded important insights into the nature of wandering and its consequences. However, neither of these approaches measures wandering directly or exclusively. All activity monitoring devices currently on the market do capture wandering behaviour, but likewise, detect walking or other motion that is not wandering. At best, these devices can generate an estimate of the amount of locomotion, which may be informative when used in conjunction with other measures. Incorporation of further technological advances into these devices, such as global positioning system (GPS) or radio-frequency identification device (RFID) capabilities, are needed to render them capable of differentiating wandering from non-wandering locomotion.Using a different approach from Martino-Saltzman’s pattern classification, Kearns and colleagues ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2010</Year><RecNum>180</RecNum><DisplayText>[15]</DisplayText><record><rec-number>180</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">180</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Nams, VO</author><author>Fozard, James L</author></authors></contributors><titles><title>Tortuosity in movement paths is related to cognitive impairment</title><secondary-title>Methods Inf Med</secondary-title></titles><periodical><full-title>Methods Inf Med</full-title></periodical><pages>592-598</pages><volume>49</volume><number>6</number><dates><year>2010</year></dates><isbn>0026-1270</isbn><urls></urls></record></Cite></EndNote>[15] used Ultra-wideband real time location systems to measure physical movement and location of dozens of ALF residents in near real-time for several months. Instead of categorizing movement paths, the amount of randomness contained in each movement path is quantified using Fractal Dimension, a dimensionless number ranging from 1 (perfectly straight) to 2 (completely random) that estimates tortuosity of the path. Any path, irrespective of source, may have its randomness quantified using Fractal Dimension. Kearns et al. ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2016</Year><RecNum>9039</RecNum><DisplayText>[16]</DisplayText><record><rec-number>9039</rec-number><foreign-keys><key app="EN" db-id="avf5v0ptmz900ne2xem5999f2tzfas2pdzsd" timestamp="1487280420">9039</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, W.D. </author><author>Fozard, J.L.</author><author>Nams, V.O.</author></authors></contributors><titles><title>Movement path tortuosity in free ambulation: Relationships to age and brain disease</title><secondary-title>IEEE Journal of Biomedical and Health Informatics</secondary-title><short-title>Movement path tortuosity in free ambulation: Relationships to age and brain disease.</short-title></titles><periodical><full-title>IEEE Journal of Biomedical and Health Informatics</full-title></periodical><volume>Published ahead of print</volume><edition>2016</edition><dates><year>2016</year><pub-dates><date>2016</date></pub-dates></dates><urls></urls><electronic-resource-num>doi:10.1109/JBHI.2016.2517332</electronic-resource-num></record></Cite></EndNote>[16] have discovered that subjects’ paths mean Fractal Dimension values correlate with cognitive test such as Mini Mental State Examination (MMSE). Fractal Dimension is also a significant predictor of impending falls ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2012</Year><RecNum>446</RecNum><DisplayText>[17]</DisplayText><record><rec-number>446</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516243941">446</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, W. D.</author><author>Fozard, J. L.</author><author>Becker, M.</author><author>Jasiewicz, J. M.</author><author>Craighead, J. D.</author><author>Holtsclaw, L.</author><author>Dion, C.</author></authors></contributors><titles><title>Path Tortuosity in Everyday Movements of Elderly Persons Increases Fall Prediction Beyond Knowledge of Fall History, Medication Use, and Standardized Gait and Balance Assessments</title><secondary-title>Journal of the American Medical Directors Association</secondary-title></titles><periodical><full-title>Journal of the American Medical Directors Association</full-title></periodical><pages>665.e7-665.e13</pages><volume>13</volume><number>7</number><dates><year>2012</year></dates><urls></urls><electronic-resource-num>10.1016/j.jamda.2012.06.010</electronic-resource-num></record></Cite></EndNote>[17]. Kearns et al.’s approach quantifies random characteristics (tortuosity) of paths generated by wanderers in near real-time, revealing the tortuosity values for each path gathered while conducting normal daily activities over time. Many more techniques and devices were developed using the technological approach which have been discussed in detail in later section.Need and opportunity to measure WanderingWandering poses serious concerns for family and professional caregivers and remains a topic of significant interest for scientists, clinicians, and policy makers because the behavior is associated with some of the gravest adverse outcomes including getting lost, falling, and an increased risk of death. Reflecting this concern, a growing number of studies [6] have focused on various aspects of this pervasive and intriguing behavior that include: definition, measurement, the outcomes of wandering, wandering-related behaviors such as intrusion, shadowing, and techniques for wandering management and intervention.Real-time location systems provide a convenient way to track a persons’ trajectory in 3D space. Trajectories are recorded as sequence of coordinates where each data point represents subject identity, location, and time stamp information [7]. Position tracking devices exist for outdoor use, such as Global Positioning System (GPS), and Received Signal Strength (RSS) can provide relatively crude indoor location fixes when combined with Radio-frequency identification (RFID) readers. More elaborate, and expensive options for position tracking using Ultra-Wideband (UWB) sensors for indoor scenarios are now available which provide precise indoor location information. With their integration into most ubiquitous devices such as mobile phones and smart watches, a large amount of inexpensively obtainable navigational data is now available for analysis. There is an opportunity to exploit these data to identify and analyse the significant features of movement trajectories of dementia patients. Cognitive and WanderingWandering correlates significantly with the cognitive factor ADDIN CSL_CITATION { "citationItems" : [ { "id" : "ITEM-1", "itemData" : { "ISSN" : "1532-5415", "author" : [ { "dropping-particle" : "", "family" : "Holtzer", "given" : "Roee", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Tang", "given" : "Ming\u2010Xin", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Devanand", "given" : "D P", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Albert", "given" : "Steven M", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Wegesin", "given" : "Domonick J", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Marder", "given" : "Karen", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Bell", "given" : "Karen", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Albert", "given" : "Marilyn", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Brandt", "given" : "Jason", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Stern", "given" : "Yaakov", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" } ], "container-title" : "Journal of the American Geriatrics Society", "id" : "ITEM-1", "issue" : "7", "issued" : { "date-parts" : [ [ "2003" ] ] }, "page" : "953-960", "publisher" : "Wiley Online Library", "title" : "Psychopathological features in Alzheimer's disease: course and relationship with cognitive status", "type" : "article-journal", "volume" : "51" }, "uris" : [ "" ] } ], "mendeley" : { "formattedCitation" : "[4]", "plainTextFormattedCitation" : "[4]", "previouslyFormattedCitation" : "[4]" }, "properties" : { "noteIndex" : 0 }, "schema" : "" }[4]. Due to the loss in cognition person lose track of place and time. The spatial and temporal disorientation aids to the onset of wandering movement. The random pattern has been associated with the loss of cognitive function ADDIN CSL_CITATION { "citationItems" : [ { "id" : "ITEM-1", "itemData" : { "ISSN" : "0193-9459", "author" : [ { "dropping-particle" : "", "family" : "Algase", "given" : "Donna L", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Beattie", "given" : "Elizabeth R A", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Therrien", "given" : "Barbara", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" } ], "container-title" : "Western Journal of Nursing Research", "id" : "ITEM-1", "issue" : "3", "issued" : { "date-parts" : [ [ "2001" ] ] }, "page" : "283-295", "publisher" : "Sage Publications", "title" : "Impact of cognitive impairment on wandering behavior", "type" : "article-journal", "volume" : "23" }, "uris" : [ "" ] } ], "mendeley" : { "formattedCitation" : "[2]", "plainTextFormattedCitation" : "[2]", "previouslyFormattedCitation" : "[2]" }, "properties" : { "noteIndex" : 0 }, "schema" : "" }[2] but still there is no enough empirical evidence to characterize lapping and pacing pattern. Lapping is less associated with the loss of cognition and generally followed by a short span of rest after the episodes. This generally occurs when the distance between target is more, these patterns vanishes once this distance is reduced. Pacing is perceived as a sign of agitation or anger and correlates very poorly with cognitive loss. It may be a considerable point of interest to know, how these patterns for a person changes over time. More research is needed to factor out relationship between wandering and cognition. Some of the most prevalent test for assessment of cognitive state of elderly are:The Mini Mental State Exam (MMSE)MMSE [16] is used for routine cognitive assessment of older adults. The MMSE is an 11-question measure that assesses five areas of cognitive function: orientation, registration, attention and calculation, recall, and language. The maximum attainable score is 30. MMSE score and cognition state is defined as shown in REF _Ref504644369 \h Table 2.1. It is administered by trained professionals and takes around 8-10 minutes to complete. Table STYLEREF 1 \s 2. SEQ Table \* ARABIC \s 1 1: MMSE and CognitionCognitionMMSE ScoreHealthy30-25Mild24-20Moderate20-13Severe<=12The Revised Algase Wandering Scale (RAWS-CV)It is a 39-question assessment tool used to quantify wandering behavior in dementia patients [17]. The items are scored from 1 ‘never or unable’ to 5 ‘always’ ( REF _Ref504644423 \h Table 2.2) with higher scores indicating more severe wandering.Table STYLEREF 1 \s 2. SEQ Table \* ARABIC \s 1 2: Scale used in RAWS-CV12345Never/UnableSeldomSometimesUsuallyAlwaysRAWS consisted of four aspects of wandering. The Community Version (CV) includes additional two aspects ( REF _Ref504644457 \h Table 2.3). The overall scores (highest possible 39 x 5 = 195) can be calculated to evaluate wandering with the highest scores indicating severe wandering behaviour. Sub-scale scores can also be calculated to evaluate the different aspects of wandering. The test is designed to be completed by staff members familiar with person with dementia. The test usually takes 10 minutes. Table STYLEREF 1 \s 2. SEQ Table \* ARABIC \s 1 3: Sub-scales of RAWS-CVAspects of wandering ?No of itemsPersistent walking14Repetitive walking7Spatial disorientation8Eloping behavior4New aspects in CV?Negative outcomes4Mealtime impulsivity2Total?39Technologies to manage wandering – A survey There have been plenty of technological solution developed over 30 years of research in this area. In this section, the author has done a systematic survey of these technological solutions. It has also been restricted to the published articles in English. Keyword search were conducted in Scopus and Google Scholar database using search term: (“wandering” OR “wandering behavior”) AND (“technology” OR “device” OR “equipment”) AND (“people with dementia” OR “people with cognitive impairment” OR “elderly”)40 articles were hand-picked for analysis, based on its relevance for our study. Three types of key techniques were applied in existing work for wandering research, each has been tabulated in a separate table:Event monitoring ( REF _Ref504644972 \h \* MERGEFORMAT Table 2.4): Using this technique sequence of events were identified to discover rhythmic repetition of event that may correspond to wandering.Trajectory tracking ( REF _Ref504644986 \h Table 2.5): Using this technique, spatiotemporal data was analyzed to find fine-grained features which can identify wandering behaviour.Localization ( REF _Ref504645001 \h Table 2.6): This technique uses GPS sensor to localize the wanderer in outdoor environment, targeting the prevention of elopements or boundary transgressions.Further in each category the devices were evaluated on the following 8 parameters as shown in REF _Ref504559679 \h Figure 2.1:Figure 2.1: Parameters used to evaluate the wandering technologySensors ModalitiesA wider variety of sensors such as GPS PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5NdWx2ZW5uYTwvQXV0aG9yPjxZZWFyPjIwMTA8L1llYXI+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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5NdWx2ZW5uYTwvQXV0aG9yPjxZZWFyPjIwMTA8L1llYXI+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ADDIN EN.CITE.DATA [18-20], Ultra-Wideband real-time location systems PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5PdGE8L0F1dGhvcj48WWVhcj4yMDExPC9ZZWFyPjxSZWNO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 EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5PdGE8L0F1dGhvcj48WWVhcj4yMDExPC9ZZWFyPjxSZWNO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 EN.CITE.DATA [21-24], Wi-Fi signal based positioning ADDIN EN.CITE <EndNote><Cite><Author>Lim</Author><Year>2007</Year><RecNum>190</RecNum><DisplayText>[25]</DisplayText><record><rec-number>190</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">190</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Lim, Chin-Heng</author><author>Wan, Yahong</author><author>Ng, Boon-Poh</author><author>See, Chong-Meng Samson</author></authors></contributors><titles><title>A real-time indoor WiFi localization system utilizing smart antennas</title><secondary-title>IEEE Transactions on Consumer Electronics</secondary-title></titles><periodical><full-title>IEEE Transactions on Consumer Electronics</full-title></periodical><volume>53</volume><number>2</number><dates><year>2007</year></dates><isbn>0098-3063</isbn><urls></urls></record></Cite></EndNote>[25], active and passive tag Radio Frequency Identification (RFID) ADDIN EN.CITE <EndNote><Cite><Author>Nugent</Author><Year>2006</Year><RecNum>191</RecNum><DisplayText>[26]</DisplayText><record><rec-number>191</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">191</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Nugent, C</author><author>Augusto, JC</author></authors></contributors><titles><title>A system for activity monitoring and patient tracking in a smart hospital</title><secondary-title>Smart Homes and Beyond: ICOST 2006: 4th International Conference on Smart Homes and Health Telematics</secondary-title></titles><pages>196</pages><volume>19</volume><dates><year>2006</year></dates><publisher>IOS Press</publisher><isbn>1586036238</isbn><urls></urls></record></Cite></EndNote>[26], or GSM sensor ADDIN EN.CITE <EndNote><Cite><Author>Miskelly</Author><Year>2005</Year><RecNum>193</RecNum><DisplayText>[27]</DisplayText><record><rec-number>193</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">193</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Miskelly, Frank</author></authors></contributors><titles><title>Electronic tracking of patients with dementia and wandering using mobile phone technology</title><secondary-title>Age and ageing</secondary-title></titles><periodical><full-title>Age and ageing</full-title></periodical><pages>497-498</pages><volume>34</volume><number>5</number><dates><year>2005</year></dates><isbn>0002-0729</isbn><urls></urls></record></Cite></EndNote>[27] can be used by researchers for developing wandering management tool. These sensors have the limitation when it comes to be used for hybrid environment (indoor and outdoor). Analyzing the methods used so far from this prospective can be useful in design of the future assistive technology. EnvironmentMany researches have been carried on in a highly controlled environment, which creates uncertainty when it comes to being used in real-world. These devices cannot scale mainly due to numerous assumptions employed during the design of the system. The author has analyzed the system from the point of view of the environment the devices have been tested and four main categories have been identified, these are:Hospitals/ Nursing home: These are long support care facility for dementia patients. PwDs are monitored around the clock. Usually, the residents of these facilities are at the severe stage of dementia and need help in carrying out ADLs.Assisted Living Facility (ALF): Distinction between nursing homes and ALF is quite blurred. But this specializes in providing long-term care alternative for older adults ADDIN EN.CITE <EndNote><Cite><Author>Smith</Author><Year>2008</Year><RecNum>522</RecNum><DisplayText>[28]</DisplayText><record><rec-number>522</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516425499">522</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Smith, Marianne</author><author>Buckwalter, Kathleen C</author><author>Kang, Hyunwook</author><author>Ellingrod, Vicki</author><author>Schultz, Susan K</author></authors></contributors><titles><title>Dementia care in assisted living: Needs and challenges</title><secondary-title>Issues in mental health nursing</secondary-title></titles><periodical><full-title>Issues in mental health nursing</full-title></periodical><pages>817-838</pages><volume>29</volume><number>8</number><dates><year>2008</year></dates><isbn>0161-2840</isbn><urls></urls></record></Cite></EndNote>[28]. Usually, the residents have moderate to severe dementia. Recently, it is becoming an increasingly popular option for dementia care. Home: These are the most autonomous system where PwD usually takes care of its health and daily activity. Usually, people who are at an early stage of dementia can enjoy their daily life without any assistance or intervention. Technology such as alert system and event monitoring is very beneficial for in this environment. Outdoor: Most of the devices developed for the outdoor aids in navigation and alert when the subject crosses any predefined geographical boundary. It generates alert to the caretakers or family members. The device used for outdoor monitoring are partial intrusive in the sense that the location of the patients is constantly monitored/tacked.Support for Independent LivingActivity of daily living (ADL) is an index which has been developed based on the methodological and observational study for twenty years ADDIN EN.CITE <EndNote><Cite><Author>Katz</Author><Year>1976</Year><RecNum>521</RecNum><DisplayText>[29]</DisplayText><record><rec-number>521</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516421126">521</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Katz, Sidney</author><author>Akpom, C Amechi</author></authors></contributors><titles><title>A measure of primary sociobiological functions</title><secondary-title>International journal of health services</secondary-title></titles><periodical><full-title>International journal of health services</full-title></periodical><pages>493-508</pages><volume>6</volume><number>3</number><dates><year>1976</year></dates><isbn>0020-7314</isbn><urls></urls></record></Cite></EndNote>[29]. It is used in profiling the severity of chronic illness based six level of sociobiological function, namely, bathing, dressing, toileting, transfer, continence, and feeding. It has contributed information about health needs and outcomes which is useful for management, planning, policy making, research, and teaching. Since the ability to perform ADLs highly depends on the cognitive state of a person and dementia being a progressive condition in which cognitive function decile over time, an assistive technology which can aid in carrying out ADLs can be immensely beneficial.Activity recognition approachThe devices used for the wandering management are also capable of monitoring the activities of the patients. Many of these devices have resorted to this kind of approach for wandering detection. The aided benefit from this approach is that it can be used in a wide range of scenarios for activity identification, suggestion, etc. The problem with such system is that it's not very accurate and are sometimes very specialized in identifying activities.InvasivenessTechnologies to manage wandering should not be invasive unless the PwD has a very high risk to harm itself. The essential dignity and privacy of the person with dementia must be always respected, and only the minimum monitoring necessary to keep them safe should be employed. Devices using cameras are considered extremely invasive. In order to make the invasion of privacy less unpalatable, monitoring devices relying on pressure transducers and switch closures and other devices have been employed instead of cameras. GPS tracking is another example where concern for invasion of privacy is debatable. Devices use it to provide tracking information to caretakers and family members. It is particularly useful for the people who are forgettable or tend to wander away from home. It can guide them to come back to home safe. The author has categorized such devices as a partially invasive.Constrain/RestrictivenessThe intervention for wandering should never be restrictive in nature. In addressing problems arising from wandering it should preserve the dignity, personhood, and individuality of the wanderer. While this goal is relevant to all wanderers, and to non-wanderers as well, it is especially important for those with problematic wandering as they are the ones for whom restrictions are most likely to be imposed. The author has tried to identify the devices which are restrictive in nature.Support for notification SystemTechnological solution for notification has been implemented by many devices in the form of alerts. Caregivers, staffs, or family members are alerted to the situation when wandering is detected. Some system has also implemented the alert system when abnormal activity is detected. It is very helpful in avoiding a serious situation. Fall detectionWanderers are at increased risk of falling and falls are the most frequently reported adverse events for residents with dementia in institutional settings. Those who fall are more likely to sustain fractures. Fall is a leading cause of death in PwD. There are several fall detectors that are currently available which are used by older adult. This range from user-activated fall detector ADDIN EN.CITE <EndNote><Cite><Author>Alwan</Author><Year>2006</Year><RecNum>541</RecNum><DisplayText>[30, 31]</DisplayText><record><rec-number>541</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1534345958">541</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Alwan, Majd</author><author>Rajendran, Prabhu Jude</author><author>Kell, Steve</author><author>Mack, David</author><author>Dalal, Siddharth</author><author>Wolfe, Matt</author><author>Felder, Robin</author></authors></contributors><titles><title>A smart and passive floor-vibration based fall detector for elderly</title><secondary-title>Information and Communication Technologies</secondary-title></titles><pages>1003-1007</pages><volume>1</volume><dates><year>2006</year></dates><urls></urls></record></Cite><Cite><Author>Sixsmith</Author><Year>2004</Year><RecNum>540</RecNum><record><rec-number>540</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1534345929">540</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Sixsmith, Andrew</author><author>Johnson, Neil</author></authors></contributors><titles><title>A smart sensor to detect the falls of the elderly</title><secondary-title>IEEE Pervasive computing</secondary-title></titles><periodical><full-title>IEEE Pervasive Computing</full-title></periodical><pages>42-47</pages><number>2</number><dates><year>2004</year></dates><isbn>1536-1268</isbn><urls></urls></record></Cite></EndNote>[30, 31] which requires manual activation of the alarm button usually integrated with pendant or wristwatch. Another class of fall detector ADDIN EN.CITE <EndNote><Cite><Author>Bharucha</Author><Year>2009</Year><RecNum>542</RecNum><DisplayText>[32, 33]</DisplayText><record><rec-number>542</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1534346198">542</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Bharucha, Ashok J</author><author>Anand, Vivek</author><author>Forlizzi, Jodi</author><author>Dew, Mary Amanda</author><author>Reynolds III, Charles F</author><author>Stevens, Scott</author><author>Wactlar, Howard</author></authors></contributors><titles><title>Intelligent assistive technology applications to dementia care: current capabilities, limitations, and future challenges</title><secondary-title>The American journal of geriatric psychiatry</secondary-title></titles><periodical><full-title>The American Journal of Geriatric Psychiatry</full-title></periodical><pages>88-104</pages><volume>17</volume><number>2</number><dates><year>2009</year></dates><isbn>1064-7481</isbn><urls></urls></record></Cite><Cite><Author>Rowe</Author><Year>2007</Year><RecNum>543</RecNum><record><rec-number>543</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1534346208">543</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Rowe, Meredeth</author><author>Lane, Stephen</author><author>Phipps, Chad</author></authors></contributors><titles><title>CareWatch: a home monitoring system for use in homes of persons with cognitive impairment</title><secondary-title>Topics in geriatric rehabilitation</secondary-title></titles><periodical><full-title>Topics in Geriatric Rehabilitation</full-title></periodical><pages>3</pages><volume>23</volume><number>1</number><dates><year>2007</year></dates><urls></urls></record></Cite></EndNote>[32, 33] uses a combination of accelerometer or tilt sensor to automatically detect the fall. This overcomes some of the drawbacks of the manual approach. But the usefulness of such devices depends on the user’s agreement to wear such devices all the time. There is also a camera-based fall detector ADDIN EN.CITE <EndNote><Cite><Author>Lee</Author><Year>2005</Year><RecNum>545</RecNum><DisplayText>[34, 35]</DisplayText><record><rec-number>545</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1534346451">545</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Lee, Tracy</author><author>Mihailidis, Alex</author></authors></contributors><titles><title>An intelligent emergency response system: preliminary development and testing of automated fall detection</title><secondary-title>Journal of telemedicine and telecare</secondary-title></titles><periodical><full-title>Journal of telemedicine and telecare</full-title></periodical><pages>194-198</pages><volume>11</volume><number>4</number><dates><year>2005</year></dates><isbn>1357-633X</isbn><urls></urls></record></Cite><Cite><Author>Nait-Charif</Author><Year>2004</Year><RecNum>546</RecNum><record><rec-number>546</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1534346469">546</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Nait-Charif, Hammadi</author><author>McKenna, Stephen J</author></authors></contributors><titles><title>Activity summarisation and fall detection in a supportive home environment</title><secondary-title>Pattern Recognition, 2004. ICPR 2004. Proceedings of the 17th International Conference on</secondary-title></titles><pages>323-326</pages><volume>4</volume><dates><year>2004</year></dates><publisher>IEEE</publisher><isbn>0769521282</isbn><urls></urls></record></Cite></EndNote>[34, 35], these devices track the resident using a camera and detects the event of fall based on the identification of unusual activity. But such devices infringe on the privacy of the user. Usually, elderly people live alone, and it is very difficult for them to report fall as soon as it happens particularly because they do not have the physical capability or become unconscious after it. Such devices can be useful for alerting the caretakes and many lives can be saved by taking timely action.Table STYLEREF 1 \s 2. SEQ Table \* ARABIC \s 1 4: Event monitoring-based Wandering detectionArticlePub YearSensor UsedLocationSupport Ind. LivingMeasures ADLInvasiveConstrain-ingNotifi-cationFall DetectionMethodIDType ADDIN EN.CITE <EndNote><Cite><Author>Doughty</Author><Year>1998</Year><RecNum>476</RecNum><DisplayText>[36]</DisplayText><record><rec-number>476</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516258849">476</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Doughty, K</author><author>Williams, G</author><author>King, PJ</author><author>Woods, R</author></authors></contributors><titles><title>DIANA-a telecare system for supporting dementia sufferers in the community</title><secondary-title>Engineering in Medicine and Biology Society, 1998. Proceedings of the 20th Annual International Conference of the IEEE</secondary-title></titles><pages>1980-1983</pages><volume>4</volume><dates><year>1998</year></dates><publisher>IEEE</publisher><isbn>0780351649</isbn><urls></urls></record></Cite></EndNote>[36]Conf.1998Piezo sensor, PIRHome YesYesNoNoNoNoEvent based monitoring for wandering detection ADDIN EN.CITE <EndNote><Cite><Author>Jit</Author><Year>2006</Year><RecNum>485</RecNum><DisplayText>[37]</DisplayText><record><rec-number>485</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">485</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Jit, B.</author><author>Zhang, D.</author><author>Qiao, G.</author><author>Foo, V.</author><author>Qiu, Q.</author><author>Yap, P.</author></authors></contributors><titles><title>A system for activity monitoring and patient tracking in a smart hospital</title><secondary-title>Proceedings of the International Conference on Smart Homes and Health Telematics, ICOST 2006</secondary-title></titles><periodical><full-title>Proceedings of the International Conference on Smart Homes and Health Telematics, ICOST 2006</full-title></periodical><pages>196-203</pages><dates><year>2006</year></dates><urls></urls></record></Cite></EndNote>[37]Jour.2006RFIDHospitalYesYesNoNoNoYesSequence matching technique was used to predict the next event ADDIN EN.CITE <EndNote><Cite><Author>Masuda</Author><Year>2002</Year><RecNum>438</RecNum><DisplayText>[38]</DisplayText><record><rec-number>438</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516243941">438</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Masuda, Y.</author><author>Yoshimura, T.</author><author>Nakajima, K.</author><author>Nambu, M.</author><author>Hayakawa, T.</author><author>Tamura, T.</author></authors></contributors><titles><title>Unconstrained monitoring of prevention of wandering the elderly</title><secondary-title>Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings</secondary-title></titles><pages>1906-1907</pages><volume>3</volume><dates><year>2002</year></dates><urls></urls></record></Cite></EndNote>[38]Conf.2002Contact sensorHospitalNoNoNoNoYesNoWandering detection based on monitoring contact with mat ADDIN EN.CITE <EndNote><Cite><Author>Ota</Author><Year>2011</Year><RecNum>437</RecNum><DisplayText>[39]</DisplayText><record><rec-number>437</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516243941">437</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Ota, K.</author><author>Ota, Y.</author><author>Otsu, M.</author><author>Kajiwara, A.</author></authors></contributors><titles><title>Elderly-care motion sensor using UWB-IR</title><secondary-title>SAS 2011 - IEEE Sensors Applications Symposium, Proceedings</secondary-title></titles><pages>159-162</pages><dates><year>2011</year></dates><urls></urls><custom7>5739786</custom7><electronic-resource-num>10.1109/SAS.2011.5739786</electronic-resource-num></record></Cite></EndNote>[39]Conf.2011UWB-IR**ALF*NoYesNoNoNoYesState detection algorithm based on ranging and motion estimation ADDIN EN.CITE <EndNote><Cite><Author>Rowe</Author><Year>2007</Year><RecNum>486</RecNum><DisplayText>[40]</DisplayText><record><rec-number>486</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">486</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Rowe, M.</author><author>Lane, S.</author><author>Phipps, C.</author></authors></contributors><titles><title>CareWatch: A home monitoring system for use in homes of persons with cognitive impairment</title><secondary-title>Topics in Geriatric Rehabilitation</secondary-title></titles><periodical><full-title>Topics in Geriatric Rehabilitation</full-title></periodical><pages>3-8</pages><volume>23</volume><number>1</number><dates><year>2007</year></dates><urls></urls><electronic-resource-num>10.1097/00013614-200701000-00003</electronic-resource-num></record></Cite></EndNote>[40]Jour.2007Air pressure bagHomeNoNoNoYesYesNoEvent based detection* Assisted Living Facility** Ultrawide Band- InfraredTable STYLEREF 1 \s 2. SEQ Table \* ARABIC \s 1 5: Trajectory tracking-based wandering detectionArticlePub YearSensor UsedLocationSupport Ind. LivingMeasures ADLInvasiveConst-rainingNotific-ationFall DetectionMethodIDType ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>480</RecNum><DisplayText>[41]</DisplayText><record><rec-number>480</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">480</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, D.</author><author>Blasch, B. B.</author><author>Morris, R. D.</author><author>McNeal, L. W.</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>Gerontologist</secondary-title></titles><periodical><full-title>Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><urls></urls><electronic-resource-num>10.1093/geront/31.5.666</electronic-resource-num></record></Cite></EndNote>[41]Jour.1991VideoNursing homeNoNoYesNoNoNoWandering Pattern detection based on observation ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2003</Year><RecNum>487</RecNum><DisplayText>[42]</DisplayText><record><rec-number>487</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">487</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, D. L.</author><author>Beattie, E. R. A.</author><author>Leitsch, S. A.</author><author>Beel-Bates, C. A.</author></authors></contributors><titles><title>Biomechanical activity devices to index wandering behavior in dementia</title><secondary-title>American Journal of Alzheimer's Disease and other Dementias</secondary-title></titles><periodical><full-title>Am J Alzheimers Dis Other Demen</full-title><abbr-1>American journal of Alzheimer's disease and other dementias</abbr-1></periodical><pages>85-92</pages><volume>18</volume><number>2</number><dates><year>2003</year></dates><urls></urls></record></Cite></EndNote>[42]Jour.2003Step Watch sensorNursing HomeNoNoNoNoNoNoCounting the number of stepsPEVuZE5vdGU+PENpdGU+PEF1dGhvcj5LZWFybnM8L0F1dGhvcj48WWVhcj4yMDA5PC9ZZWFyPjxS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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5LZWFybnM8L0F1dGhvcj48WWVhcj4yMDA5PC9ZZWFyPjxS
ZWNOdW0+NDkwPC9SZWNOdW0+PERpc3BsYXlUZXh0PlsxNywgNDMtNDddPC9EaXNwbGF5VGV4dD48
cmVjb3JkPjxyZWMtbnVtYmVyPjQ5MDwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBw
PSJFTiIgZGItaWQ9IjVhZnYwcDJkc3dldHRtZXRzMjVwZTJwZzJmenQyOTk5OWVmdiIgdGltZXN0
YW1wPSIxNTE2MjYwODUwIj40OTA8L2tleT48L2ZvcmVpZ24ta2V5cz48cmVmLXR5cGUgbmFtZT0i
Sm91cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNvbnRyaWJ1dG9ycz48YXV0aG9ycz48YXV0
aG9yPktlYXJucywgVy48L2F1dGhvcj48YXV0aG9yPkZvemFyZCwgSi4gTC48L2F1dGhvcj48L2F1
dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxlcz48dGl0bGU+RXZhbHVhdGlvbiBvZiBXYW5kZXJp
bmcgYnkgUmVzaWRlbnRzIGluIGFuIEFzc2lzdGVkIExpdmluZyBGYWNpbGl0eSAoQUxGKSB1c2lu
ZyBVbHRyYSBXaWRlYmFuZCBSYWRpbyBSVExTPC90aXRsZT48c2Vjb25kYXJ5LXRpdGxlPkogTnV0
cml0aW9uIEhlYWx0aCBBZ2luZzwvc2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxwZXJpb2RpY2Fs
PjxmdWxsLXRpdGxlPkogTnV0cml0aW9uIEhlYWx0aCBBZ2luZzwvZnVsbC10aXRsZT48L3Blcmlv
ZGljYWw+PHZvbHVtZT4xMzwvdm9sdW1lPjxkYXRlcz48eWVhcj4yMDA5PC95ZWFyPjwvZGF0ZXM+
PHVybHM+PC91cmxzPjwvcmVjb3JkPjwvQ2l0ZT48Q2l0ZT48QXV0aG9yPktlYXJuczwvQXV0aG9y
PjxZZWFyPjIwMDg8L1llYXI+PFJlY051bT40ODk8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVy
PjQ4OTwvcmVjLW51bWJlcj48Zm9yZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9IjVhZnYw
cDJkc3dldHRtZXRzMjVwZTJwZzJmenQyOTk5OWVmdiIgdGltZXN0YW1wPSIxNTE2MjYwODUwIj40
ODk8L2tleT48L2ZvcmVpZ24ta2V5cz48cmVmLXR5cGUgbmFtZT0iSm91cm5hbCBBcnRpY2xlIj4x
NzwvcmVmLXR5cGU+PGNvbnRyaWJ1dG9ycz48YXV0aG9ycz48YXV0aG9yPktlYXJucywgVy4gRC48
L2F1dGhvcj48YXV0aG9yPkFsZ2FzZSwgRC48L2F1dGhvcj48YXV0aG9yPk1vb3JlLCBELiBILjwv
YXV0aG9yPjxhdXRob3I+QWhtZWQsIFMuPC9hdXRob3I+PC9hdXRob3JzPjwvY29udHJpYnV0b3Jz
Pjx0aXRsZXM+PHRpdGxlPlVsdHJhIHdpZGViYW5kIHJhZGlvOiBBIG5vdmVsIG1ldGhvZCBmb3Ig
bWVhc3VyaW5nIHdhbmRlcmluZyBpbiBwZXJzb25zIHdpdGggZGVtZW50aWE8L3RpdGxlPjxzZWNv
bmRhcnktdGl0bGU+R2Vyb250ZWNobm9sb2d5PC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBl
cmlvZGljYWw+PGZ1bGwtdGl0bGU+R2Vyb250ZWNobm9sb2d5PC9mdWxsLXRpdGxlPjwvcGVyaW9k
aWNhbD48cGFnZXM+NDgtNTc8L3BhZ2VzPjx2b2x1bWU+Nzwvdm9sdW1lPjxudW1iZXI+MTwvbnVt
YmVyPjxkYXRlcz48eWVhcj4yMDA4PC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91cmxzPjwvcmVjb3Jk
PjwvQ2l0ZT48Q2l0ZT48QXV0aG9yPktlYXJuczwvQXV0aG9yPjxZZWFyPjIwMTI8L1llYXI+PFJl
Y051bT40OTI8L1JlY051bT48cmVjb3JkPjxyZWMtbnVtYmVyPjQ5MjwvcmVjLW51bWJlcj48Zm9y
ZWlnbi1rZXlzPjxrZXkgYXBwPSJFTiIgZGItaWQ9IjVhZnYwcDJkc3dldHRtZXRzMjVwZTJwZzJm
enQyOTk5OWVmdiIgdGltZXN0YW1wPSIxNTE2MjYwODUwIj40OTI8L2tleT48L2ZvcmVpZ24ta2V5
cz48cmVmLXR5cGUgbmFtZT0iSm91cm5hbCBBcnRpY2xlIj4xNzwvcmVmLXR5cGU+PGNvbnRyaWJ1
dG9ycz48YXV0aG9ycz48YXV0aG9yPktlYXJucywgVy4gRC48L2F1dGhvcj48YXV0aG9yPkZvemFy
ZCwgSi4gTC48L2F1dGhvcj48YXV0aG9yPkJlY2tlciwgTS48L2F1dGhvcj48YXV0aG9yPkphc2ll
d2ljeiwgSi4gTS48L2F1dGhvcj48YXV0aG9yPkNyYWlnaGVhZCwgSi4gRC48L2F1dGhvcj48YXV0
aG9yPkhvbHRzY2xhdywgTC48L2F1dGhvcj48YXV0aG9yPkRpb24sIEMuPC9hdXRob3I+PC9hdXRo
b3JzPjwvY29udHJpYnV0b3JzPjx0aXRsZXM+PHRpdGxlPlBhdGggVG9ydHVvc2l0eSBpbiBFdmVy
eWRheSBNb3ZlbWVudHMgb2YgRWxkZXJseSBQZXJzb25zIEluY3JlYXNlcyBGYWxsIFByZWRpY3Rp
b24gQmV5b25kIEtub3dsZWRnZSBvZiBGYWxsIEhpc3RvcnksIE1lZGljYXRpb24gVXNlLCBhbmQg
U3RhbmRhcmRpemVkIEdhaXQgYW5kIEJhbGFuY2UgQXNzZXNzbWVudHM8L3RpdGxlPjxzZWNvbmRh
cnktdGl0bGU+Sm91cm5hbCBvZiB0aGUgQW1lcmljYW4gTWVkaWNhbCBEaXJlY3RvcnMgQXNzb2Np
YXRpb248L3NlY29uZGFyeS10aXRsZT48L3RpdGxlcz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5K
b3VybmFsIG9mIHRoZSBBbWVyaWNhbiBNZWRpY2FsIERpcmVjdG9ycyBBc3NvY2lhdGlvbjwvZnVs
bC10aXRsZT48L3BlcmlvZGljYWw+PHBhZ2VzPjY2NS5lNy02NjUuZTEzPC9wYWdlcz48dm9sdW1l
PjEzPC92b2x1bWU+PG51bWJlcj43PC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTI8L3llYXI+PC9k
YXRlcz48dXJscz48L3VybHM+PGVsZWN0cm9uaWMtcmVzb3VyY2UtbnVtPjEwLjEwMTYvai5qYW1k
YS4yMDEyLjA2LjAxMDwvZWxlY3Ryb25pYy1yZXNvdXJjZS1udW0+PC9yZWNvcmQ+PC9DaXRlPjxD
aXRlPjxBdXRob3I+S2Vhcm5zPC9BdXRob3I+PFllYXI+MjAxMTwvWWVhcj48UmVjTnVtPjQ5MTwv
UmVjTnVtPjxyZWNvcmQ+PHJlYy1udW1iZXI+NDkxPC9yZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+
PGtleSBhcHA9IkVOIiBkYi1pZD0iNWFmdjBwMmRzd2V0dG1ldHMyNXBlMnBnMmZ6dDI5OTk5ZWZ2
IiB0aW1lc3RhbXA9IjE1MTYyNjA4NTAiPjQ5MTwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlw
ZSBuYW1lPSJKb3VybmFsIEFydGljbGUiPjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRo
b3JzPjxhdXRob3I+S2Vhcm5zLCBXLiBELjwvYXV0aG9yPjxhdXRob3I+Rm96YXJkLCBKLiBMLjwv
YXV0aG9yPjxhdXRob3I+TmFtcywgVi4gTy48L2F1dGhvcj48YXV0aG9yPkNyYWlnaGVhZCwgSi4g
RC48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmlidXRvcnM+PHRpdGxlcz48dGl0bGU+V2lyZWxl
c3MgdGVsZXN1cnZlaWxsYW5jZSBzeXN0ZW0gZm9yIGRldGVjdGluZyBkZW1lbnRpYTwvdGl0bGU+
PHNlY29uZGFyeS10aXRsZT5HZXJvbnRlY2hub2xvZ3k8L3NlY29uZGFyeS10aXRsZT48L3RpdGxl
cz48cGVyaW9kaWNhbD48ZnVsbC10aXRsZT5HZXJvbnRlY2hub2xvZ3k8L2Z1bGwtdGl0bGU+PC9w
ZXJpb2RpY2FsPjxwYWdlcz45MC0xMDI8L3BhZ2VzPjx2b2x1bWU+MTA8L3ZvbHVtZT48bnVtYmVy
PjI8L251bWJlcj48ZGF0ZXM+PHllYXI+MjAxMTwveWVhcj48L2RhdGVzPjx1cmxzPjwvdXJscz48
ZWxlY3Ryb25pYy1yZXNvdXJjZS1udW0+MTAuNDAxNy9ndC4yMDExLjEwLjIuMDA0LjAwPC9lbGVj
dHJvbmljLXJlc291cmNlLW51bT48L3JlY29yZD48L0NpdGU+PENpdGU+PEF1dGhvcj5LZWFybnM8
L0F1dGhvcj48WWVhcj4yMDEwPC9ZZWFyPjxSZWNOdW0+NDg4PC9SZWNOdW0+PHJlY29yZD48cmVj
LW51bWJlcj40ODg8L3JlYy1udW1iZXI+PGZvcmVpZ24ta2V5cz48a2V5IGFwcD0iRU4iIGRiLWlk
PSI1YWZ2MHAyZHN3ZXR0bWV0czI1cGUycGcyZnp0Mjk5OTllZnYiIHRpbWVzdGFtcD0iMTUxNjI2
MDg1MCI+NDg4PC9rZXk+PC9mb3JlaWduLWtleXM+PHJlZi10eXBlIG5hbWU9IkpvdXJuYWwgQXJ0
aWNsZSI+MTc8L3JlZi10eXBlPjxjb250cmlidXRvcnM+PGF1dGhvcnM+PGF1dGhvcj5LZWFybnMs
IFcuIEQuPC9hdXRob3I+PGF1dGhvcj5OYW1zLCBWLiBPLjwvYXV0aG9yPjxhdXRob3I+Rm96YXJk
LCBKLiBMLjwvYXV0aG9yPjwvYXV0aG9ycz48L2NvbnRyaWJ1dG9ycz48dGl0bGVzPjx0aXRsZT5U
b3J0dW9zaXR5IGluIG1vdmVtZW50IHBhdGhzIGlzIHJlbGF0ZWQgdG8gY29nbml0aXZlIGltcGFp
cm1lbnQgd2lyZWxlc3MgZnJhY3RhbCBFc3RpbWF0aW9uIGluIGFzc2lzdGVkIGxpdmluZyBmYWNp
bGl0eSByZXNpZGVudHM8L3RpdGxlPjxzZWNvbmRhcnktdGl0bGU+TWV0aG9kcyBvZiBJbmZvcm1h
dGlvbiBpbiBNZWRpY2luZTwvc2Vjb25kYXJ5LXRpdGxlPjwvdGl0bGVzPjxwZXJpb2RpY2FsPjxm
dWxsLXRpdGxlPk1ldGhvZHMgb2YgSW5mb3JtYXRpb24gaW4gTWVkaWNpbmU8L2Z1bGwtdGl0bGU+
PC9wZXJpb2RpY2FsPjxwYWdlcz41OTItNTk4PC9wYWdlcz48dm9sdW1lPjQ5PC92b2x1bWU+PG51
bWJlcj42PC9udW1iZXI+PGRhdGVzPjx5ZWFyPjIwMTA8L3llYXI+PC9kYXRlcz48dXJscz48L3Vy
bHM+PGVsZWN0cm9uaWMtcmVzb3VyY2UtbnVtPjEwLjM0MTQvTUUwOS0wMS0wMDc5PC9lbGVjdHJv
bmljLXJlc291cmNlLW51bT48L3JlY29yZD48L0NpdGU+PENpdGU+PEF1dGhvcj5OYW1zPC9BdXRo
b3I+PFllYXI+MjAwNDwvWWVhcj48UmVjTnVtPjUyNDwvUmVjTnVtPjxyZWNvcmQ+PHJlYy1udW1i
ZXI+NTI0PC9yZWMtbnVtYmVyPjxmb3JlaWduLWtleXM+PGtleSBhcHA9IkVOIiBkYi1pZD0iNWFm
djBwMmRzd2V0dG1ldHMyNXBlMnBnMmZ6dDI5OTk5ZWZ2IiB0aW1lc3RhbXA9IjE1MTY0NDU1MzIi
PjUyNDwva2V5PjwvZm9yZWlnbi1rZXlzPjxyZWYtdHlwZSBuYW1lPSJKb3VybmFsIEFydGljbGUi
PjE3PC9yZWYtdHlwZT48Y29udHJpYnV0b3JzPjxhdXRob3JzPjxhdXRob3I+TmFtcywgVi4gTy48
L2F1dGhvcj48YXV0aG9yPkJvdXJnZW9pcywgTS48L2F1dGhvcj48L2F1dGhvcnM+PC9jb250cmli
dXRvcnM+PHRpdGxlcz48dGl0bGU+RnJhY3RhbCBhbmFseXNpcyBtZWFzdXJlcyBoYWJpdGF0IHVz
ZSBhdCBkaWZmZXJlbnQgc3BhdGlhbCBzY2FsZXM6IEFuIGV4YW1wbGUgd2l0aCBBbWVyaWNhbiBt
YXJ0ZW48L3RpdGxlPjxzZWNvbmRhcnktdGl0bGU+Q2FuYWRpYW4gSm91cm5hbCBvZiBab29sb2d5
PC9zZWNvbmRhcnktdGl0bGU+PC90aXRsZXM+PHBlcmlvZGljYWw+PGZ1bGwtdGl0bGU+Q2FuYWRp
YW4gSm91cm5hbCBvZiBab29sb2d5PC9mdWxsLXRpdGxlPjwvcGVyaW9kaWNhbD48cGFnZXM+MTcz
OC0xNzQ3PC9wYWdlcz48dm9sdW1lPjgyPC92b2x1bWU+PG51bWJlcj4xMTwvbnVtYmVyPjxkYXRl
cz48eWVhcj4yMDA0PC95ZWFyPjwvZGF0ZXM+PHVybHM+PC91cmxzPjxlbGVjdHJvbmljLXJlc291
cmNlLW51bT4xMC4xMTM5L1owNC0xNjc8L2VsZWN0cm9uaWMtcmVzb3VyY2UtbnVtPjwvcmVjb3Jk
PjwvQ2l0ZT48L0VuZE5vdGU+
ADDIN EN.CITE.DATA [17, 43-47]Jour.2008-12RFIDHome, ALF*NoNoNoNoNoNoPath tortuosity to measure wandering ADDIN EN.CITE <EndNote><Cite><Author>Taub</Author><Year>2011</Year><RecNum>481</RecNum><DisplayText>[48]</DisplayText><record><rec-number>481</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">481</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Taub, D. M.</author><author>Lupton, E. C.</author><author>Hinman, R. T.</author><author>Leeb, S. B.</author><author>Zeisel, J.</author><author>Blackler, S.</author></authors></contributors><titles><title>The Escort system: A safety monitor for people living with Alzheimer's disease</title><secondary-title>IEEE Pervasive Computing</secondary-title></titles><periodical><full-title>IEEE Pervasive Computing</full-title></periodical><pages>68-76</pages><volume>10</volume><number>2</number><dates><year>2011</year></dates><urls></urls><custom7>5445073</custom7><electronic-resource-num>10.1109/MPRV.2010.44</electronic-resource-num></record></Cite></EndNote>[48]Jour.2011Optical SensorOutdoorNoNoPartialNoYesNoWandering movements based on the geographical patterns ADDIN EN.CITE <EndNote><Cite><Author>Vuong</Author><Year>2011</Year><RecNum>493</RecNum><DisplayText>[49]</DisplayText><record><rec-number>493</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">493</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Vuong, N. K.</author><author>Chan, S.</author><author>Lau, C. T.</author></authors></contributors><titles><title>Feasibility study of a real-time wandering detection algorithm for dementia patients</title><secondary-title>Proc Mobile Health 2011</secondary-title></titles><periodical><full-title>Proc Mobile Health 2011</full-title></periodical><dates><year>2011</year></dates><urls></urls></record></Cite></EndNote>[49]Jour.2011RFIDHomeNoNoYesNoNoNoMachine learning approach ADDIN EN.CITE <EndNote><Cite><Author>Lin</Author><Year>2012</Year><RecNum>494</RecNum><DisplayText>[50]</DisplayText><record><rec-number>494</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">494</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Lin, Q.</author><author>Zhang, D.</author><author>Huang, X.</author><author>Ni, H.</author><author>Zhou, X.</author></authors></contributors><titles><title>Detecting wandering behavior based on GPS traces for elders with dementia</title><secondary-title>2012 12th International Conference on Control, Automation, Robotics and Vision, ICARCV 2012</secondary-title></titles><pages>672-677</pages><dates><year>2012</year></dates><urls></urls><custom7>6485238</custom7><electronic-resource-num>10.1109/ICARCV.2012.6485238</electronic-resource-num></record></Cite></EndNote>[50]Conf.2012GPSOutdoorNoNoPartialNoNoNoDetection of lapping and pacing based on angle turn ADDIN EN.CITE <EndNote><Cite><Author>Liao</Author><Year>2004</Year><RecNum>523</RecNum><DisplayText>[51]</DisplayText><record><rec-number>523</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516444651">523</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Liao, L.</author><author>Fox, D.</author><author>Kautz, H.</author></authors></contributors><titles><title>Learning and inferring transportation routines</title><secondary-title>Proceedings of the National Conference on Artificial Intelligence</secondary-title></titles><pages>348-353</pages><dates><year>2004</year></dates><work-type>Conference Paper</work-type><urls><related-urls><url>;[51]Conf.2004BiomechanicalOutdoorYesNoNoNoYesNoBased on anomaly detection ADDIN EN.CITE <EndNote><Cite><Author>Chang</Author><Year>2010</Year><RecNum>496</RecNum><DisplayText>[52, 53]</DisplayText><record><rec-number>496</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">496</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Chang, Y. J.</author></authors></contributors><titles><title>Anomaly detection for travelling individuals with cognitive impairments</title><secondary-title>SIGACCESS Newsletter</secondary-title></titles><periodical><full-title>SIGACCESS Newsletter</full-title></periodical><pages>25-32</pages><volume>97</volume><dates><year>2010</year></dates><urls></urls></record></Cite><Cite><Author>Chang</Author><Year>2010</Year><RecNum>497</RecNum><record><rec-number>497</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">497</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Chang, Y. J.</author><author>Wang, T. Y.</author></authors></contributors><titles><title>Mobile location-based social networking in supported employment for people with cognitive impairments</title><secondary-title>Cybernetics and Systems</secondary-title></titles><periodical><full-title>Cybernetics and Systems</full-title></periodical><pages>245-261</pages><volume>41</volume><number>3</number><dates><year>2010</year></dates><urls></urls><electronic-resource-num>10.1080/01969721003684964</electronic-resource-num></record></Cite></EndNote>[52, 53]Jour.2010GPSOutdoorYesNoPartialNoYesNoTrajectory Analysis ADDIN EN.CITE <EndNote><Cite><Author>Lim</Author><Year>2007</Year><RecNum>498</RecNum><DisplayText>[54]</DisplayText><record><rec-number>498</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">498</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Lim, C. H.</author><author>Wan, Y.</author><author>Ng, B. P.</author><author>See, C. M. S.</author></authors></contributors><titles><title>A real-time indoor WiFi localization system utilizing smart antennas</title><secondary-title>IEEE Transactions on Consumer Electronics</secondary-title></titles><periodical><full-title>IEEE Transactions on Consumer Electronics</full-title></periodical><pages>618-622</pages><volume>53</volume><number>2</number><dates><year>2007</year></dates><urls></urls><electronic-resource-num>10.1109/TCE.2007.381737</electronic-resource-num></record></Cite></EndNote>[54]Jour.2007UWB-RTLS**HomeYesNoNoNoYesYes, OfflinePath tortuosity to measure wandering ADDIN EN.CITE <EndNote><Cite><Author>Chen</Author><Year>2007</Year><RecNum>499</RecNum><DisplayText>[55]</DisplayText><record><rec-number>499</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">499</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Chen, D.</author><author>Bharucha, A. J.</author><author>Wactlar, H. D.</author></authors></contributors><titles><title>Intelligent video monitoring to improve safety of older persons</title><secondary-title>Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings</secondary-title></titles><pages>3814-3817</pages><dates><year>2007</year></dates><urls></urls><custom7>4353163</custom7><electronic-resource-num>10.1109/IEMBS.2007.4353163</electronic-resource-num></record></Cite></EndNote>[55]Conf.2007CameraHomeYesYesYesNoYesyesComputer vision elopement detector?* Assisted Living Facility** Ultrawide Band - Real Time Location SystemTable STYLEREF 1 \s 2. SEQ Table \* ARABIC \s 1 6: Location-based prevention of wandering-related adverse eventsArticlePub YearSensor UsedLocationSupport Ind. LivingMeasures ADLInvasiveConst-rainingNotifi-cationFall DetectionMethodIDType ADDIN EN.CITE <EndNote><Cite><Author>Vuong</Author><Year>2011</Year><RecNum>501</RecNum><DisplayText>[56]</DisplayText><record><rec-number>501</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">501</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Vuong, N. K.</author><author>Chan, S.</author><author>Lau, C. T.</author><author>Lau, K. M.</author></authors></contributors><titles><title>A predictive location-aware algorithm for dementia care</title><secondary-title>Proceedings of the International Symposium on Consumer Electronics, ISCE</secondary-title></titles><pages>339-342</pages><dates><year>2011</year></dates><urls></urls><custom7>5973845</custom7><electronic-resource-num>10.1109/ISCE.2011.5973845</electronic-resource-num></record></Cite></EndNote>[56]Conf.2011GPS/cell-towerOutdoorYesNoPartialNoNoNoMachine learning approach ADDIN EN.CITE <EndNote><Cite><Author>Sposaro</Author><Year>2010</Year><RecNum>502</RecNum><DisplayText>[57]</DisplayText><record><rec-number>502</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">502</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Sposaro, F.</author><author>Danielson, J.</author><author>Tyson, G.</author></authors></contributors><titles><title>IWander: An Android application for dementia patients</title><secondary-title>2010 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC'10</secondary-title></titles><pages>3875-3878</pages><dates><year>2010</year></dates><urls></urls><custom7>5627669</custom7><electronic-resource-num>10.1109/IEMBS.2010.5627669</electronic-resource-num></record></Cite></EndNote>[57]Conf.2010Phone GPSOutdoorYesNoPartialYesYesNoProbability estimation based on mobile sensor data ADDIN EN.CITE <EndNote><Cite><Author>Wan</Author><Year>2010</Year><RecNum>504</RecNum><DisplayText>[58, 59]</DisplayText><record><rec-number>504</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">504</key></foreign-keys><ref-type name="Serial">57</ref-type><contributors><authors><author>Wan, J.</author><author>Byrne, C.</author><author>O'Hare, G. M. P.</author><author>O'Grady, M. J.</author></authors></contributors><titles><title>OutCare: Supporting dementia patients in outdoor scenarios</title><secondary-title>Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)</secondary-title></titles><pages>365-374</pages><volume>6279 LNAI</volume><dates><year>2010</year></dates><urls></urls><electronic-resource-num>10.1007/978-3-642-15384-6_39</electronic-resource-num></record></Cite><Cite><Author>Wan</Author><Year>2011</Year><RecNum>503</RecNum><record><rec-number>503</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">503</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Wan, J.</author><author>Byrne, C.</author><author>O'Hare, G. M. P.</author><author>O'Grady, M. J.</author></authors></contributors><titles><title>Orange alerts: Lessons from an outdoor case study</title><secondary-title>2011 5th International Conference on Pervasive Computing Technologies for Healthcare and Workshops, PervasiveHealth 2011</secondary-title></titles><pages>446-451</pages><dates><year>2011</year></dates><urls></urls><custom7>6038846</custom7><electronic-resource-num>10.4108/icst.pervasivehealth.2011.246081</electronic-resource-num></record></Cite></EndNote>[58, 59]Conf.2010-11GPSOutdoorYesYesPartialNoYesNoAnalyzing deviations from daily behavior signature ADDIN EN.CITE <EndNote><Cite><Author>Hoey</Author><Year>0000</Year><RecNum>506</RecNum><DisplayText>[60, 61]</DisplayText><record><rec-number>506</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">506</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Hoey, J.</author><author>Yang, X.</author><author>Favela, J.</author></authors></contributors><titles><title>Decision theoretic, context aware safety assistance for persons who wander</title><secondary-title>Proceedings of the the 7th International Workshop on Ubiquitous Health and Wellness (UbiHealth 2012)</secondary-title></titles><periodical><full-title>Proceedings of the the 7th International Workshop on Ubiquitous Health and Wellness (UbiHealth 2012)</full-title></periodical><dates><year>0000</year></dates><urls></urls></record></Cite><Cite><Author>Hoey</Author><Year>2012</Year><RecNum>505</RecNum><record><rec-number>505</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">505</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Hoey, J.</author><author>Yang, X.</author><author>Quintana, E.</author><author>Favela, J.</author></authors></contributors><titles><title>LaCasa: Location and context-aware safety assistant</title><secondary-title>2012 6th International Conference on Pervasive Computing Technologies for Healthcare and Workshops, PervasiveHealth 2012</secondary-title></titles><pages>171-174</pages><dates><year>2012</year></dates><urls></urls><electronic-resource-num>10.4108/icst.pervasivehealth.2012.248642</electronic-resource-num></record></Cite></EndNote>[60, 61]Jour., Conf.2012GPSNoBackground and contextual factors to detect risky wandering behavior ADDIN EN.CITE <EndNote><Cite><Author>Ogawa</Author><Year>2004</Year><RecNum>508</RecNum><DisplayText>[62, 63]</DisplayText><record><rec-number>508</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">508</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Ogawa, H.</author><author>Yonezawa, Y.</author><author>Maki, H.</author><author>Sato, H.</author><author>Caldwell, W. M.</author></authors></contributors><titles><title>A mobile phone-based safety support system for wandering elderly persons</title><secondary-title>Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings</secondary-title></titles><pages>3316-3317</pages><volume>26 V</volume><dates><year>2004</year></dates><urls></urls></record></Cite><Cite><Author>Rodríguez</Author><Year>2012</Year><RecNum>507</RecNum><record><rec-number>507</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">507</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Rodríguez, M. D.</author><author>Navarro, R. F.</author><author>Favela, J.</author><author>Hoey, J.</author></authors></contributors><titles><title>An ontological representation model to tailor ambient assisted interventions for wandering</title><secondary-title>AAAI Fall Symposium - Technical Report</secondary-title></titles><pages>32-37</pages><volume>FS-12-01</volume><dates><year>2012</year></dates><urls></urls></record></Cite></EndNote>[62, 63]Conf.2004, 2012GPSOutdoorYesYesPartialYesYesNoDetect risky behaviour ADDIN EN.CITE <EndNote><Cite><Author>Miskelly</Author><Year>2005</Year><RecNum>510</RecNum><DisplayText>[64, 65]</DisplayText><record><rec-number>510</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">510</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Miskelly, F.</author></authors></contributors><titles><title>Electronic tracking of patients with dementia and wandering using mobile phone technology [1]</title><secondary-title>Age and Ageing</secondary-title></titles><periodical><full-title>Age and ageing</full-title></periodical><pages>497-499</pages><volume>34</volume><number>5</number><dates><year>2005</year></dates><urls></urls><electronic-resource-num>10.1093/ageing/afi145</electronic-resource-num></record></Cite><Cite><Author>Shimizu</Author><Year>2000</Year><RecNum>511</RecNum><record><rec-number>511</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">511</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Shimizu, K.</author><author>Kawamura, K.</author><author>Yamamoto, K.</author></authors></contributors><titles><title>Location system for dementia wandering</title><secondary-title>Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings</secondary-title></titles><pages>1556-1559</pages><volume>2</volume><dates><year>2000</year></dates><urls></urls></record></Cite></EndNote>[64, 65]Jour., Conf.2005, 2000GPSOutdoorYesNoPartialNoYesNoLocalization method using GPSPEVuZE5vdGU+PENpdGU+PEF1dGhvcj5DYWx2by1QYWxvbWlubzwvQXV0aG9yPjxZZWFyPjIwMDk8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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5DYWx2by1QYWxvbWlubzwvQXV0aG9yPjxZZWFyPjIwMDk8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ADDIN EN.CITE.DATA [66-68]Jour., Conf.2006-10GPS, WIFI, RFIDHome/OutdoorYesYesNoNoYesNoPwD can send distress signal to caretakers ADDIN EN.CITE <EndNote><Cite><Year>0000</Year><RecNum>526</RecNum><DisplayText>[69, 70]</DisplayText><record><rec-number>526</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516445532">526</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors></contributors><titles><secondary-title>Wherify Wireless</secondary-title></titles><periodical><full-title>Wherify Wireless</full-title></periodical><dates><year>0000</year></dates><urls></urls></record></Cite><Cite><Author>Parnes</Author><Year>0000</Year><RecNum>525</RecNum><record><rec-number>525</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516445532">525</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Parnes, P. B.</author></authors></contributors><titles><secondary-title>GPS Technology and Alzheimer's Disease: Novel Use for An Existing Technology</secondary-title></titles><periodical><full-title>GPS Technology and Alzheimer's Disease: Novel Use for An Existing Technology</full-title></periodical><dates><year>0000</year></dates><urls></urls></record></Cite></EndNote>[69, 70]Web, Jour.1998GPSOutdoorYesNoPartialYesYesNoWandering detection method ADDIN EN.CITE <EndNote><Cite><Author>Miskelly</Author><Year>2004</Year><RecNum>515</RecNum><DisplayText>[71, 72]</DisplayText><record><rec-number>515</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516260850">515</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Miskelly, F.</author></authors></contributors><titles><title>A novel system of electronic tagging in patients with dementia and wandering</title><secondary-title>Age and Ageing</secondary-title></titles><periodical><full-title>Age and ageing</full-title></periodical><pages>304-306</pages><volume>33</volume><number>3</number><dates><year>2004</year></dates><urls></urls><electronic-resource-num>10.1093/ageing/afh084</electronic-resource-num></record></Cite><Cite><Author>Blackburn</Author><Year>1988</Year><RecNum>527</RecNum><record><rec-number>527</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516445532">527</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Blackburn, P.</author></authors></contributors><titles><title>Freedom to wander</title><secondary-title>Nursing times</secondary-title></titles><periodical><full-title>Nursing times</full-title></periodical><pages>54-55</pages><volume>84</volume><number>49</number><dates><year>1988</year></dates><urls></urls></record></Cite></EndNote>[71, 72]Jour.1988, 2004RF, Electronic taggingOutdoorNoNoYesYesYesNoBoundary alarms or electronic tagging with bracelets and monitoring PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BUERNPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BUERNPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE.DATA [70, 73-75]Device---GPSOutdoorYesNoYesYesNoNoLocalization method using GPSA study of few important devicesIn this section, the author gives a broad overview of some of the selected devices or new approach to the wandering detection which have shaped this research.Activity recognition approach for wandering detectionSang-Ho et al. ADDIN EN.CITE <EndNote><Cite><Author>Kim</Author><Year>2009</Year><RecNum>100</RecNum><DisplayText>[76]</DisplayText><record><rec-number>100</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">100</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Kim, Kyu-Jin</author><author>Hassan, Mohammad Mehedi</author><author>Na, Sang-Ho</author><author>Huh, Eui-Nam</author></authors></contributors><titles><title>Dementia wandering detection and activity recognition algorithm using tri-axial accelerometer sensors</title><secondary-title>Ubiquitous Information Technologies & Applications, 2009. ICUT'09. Proceedings of the 4th International Conference on</secondary-title></titles><pages>1-5</pages><dates><year>2009</year></dates><publisher>IEEE</publisher><isbn>1424451310</isbn><urls></urls></record></Cite></EndNote>[76] have done a systemic study of dementia wandering. They have modelled it as activity recognition (AR) with stepwise implementation of the location tracking system, modelling human mobility as Levi-walk and finally the implementation of the proposed algorithm for dementia patients. Tri-axis accelerometer and ultrasonic sensor devices were used for recording the movement. The raw sensor data from the accelerometer was converted into readable form, and then classified into the wandering/non-wandering episode based on the comparison with the previously stored patterns. They used the method to digitize the human activity and applied levy-walk model to detect wandering pattern. Levy-walk/flight is characterized by the movement pattern in animal where various small steps are taken randomly followed by long steps. Recently, researchers have reported that human walking patterns in an outdoor setting over tens of kilometers resemble a truncated form of Levy-walks, which are commonly observed in animals such as monkeys, birds and jackals. The authors have hypothesized that patient who showed wandering behaviour will show random waypoint model, whereas non-wanderer will show levy-walk model. It was based on the idea that dementia with memory loss would induce more frequent movement without any purpose.Wandering detection in indoor as well as outdoor environmentWandering can occur in an indoor as well as in an outdoor environment. N.K. Vuong et al. ADDIN EN.CITE <EndNote><Cite><Author>Vuong</Author><Year>2013</Year><RecNum>101</RecNum><DisplayText>[77]</DisplayText><record><rec-number>101</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">101</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Vuong, NK</author><author>Goh, SGA</author><author>Chan, S</author><author>Lau, CT</author></authors></contributors><titles><title>A mobile-health application to detect wandering patterns of elderly people in home environment</title><secondary-title>Engineering in Medicine and Biology Society (EMBC), 2013 35th Annual International Conference of the IEEE</secondary-title></titles><pages>6748-6751</pages><dates><year>2013</year></dates><publisher>IEEE</publisher><isbn>1557-170X</isbn><urls></urls></record></Cite></EndNote>[77] have developed mobile application that used Wi-Fi and mobile sensors to identify wandering in an indoor setup. Although this system was not very accurate in localization, but it was very cost effective as the only hardware used was mobile and Wi-Fi router. The system was also very scalable and non-obstructive. The patients were not required to carry any extra devices. Alert system was also implemented by sending SMS to caregivers as soon as wandering is detected. The system was able to classify most of the wandering patterns, but there was some classification error with random pattern, some of these were wrongly identified as direct pattern. This was the first attempt to classify the wandering pattern according to Martino-Saltzman proposed method.In another paper N.K. Vuong et al. ADDIN EN.CITE <EndNote><Cite><Author>Vuong</Author><Year>2014</Year><RecNum>102</RecNum><DisplayText>[78]</DisplayText><record><rec-number>102</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">102</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Vuong, NK</author><author>Chan, S</author><author>Lau, CT</author></authors></contributors><titles><title>Automated detection of wandering patterns in people with dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>127-147</pages><volume>12</volume><number>3</number><dates><year>2014</year></dates><isbn>1569-111X</isbn><urls></urls></record></Cite></EndNote>[78] carried on an exhaustive study for automatic classification of wandering pattern as one of the four patterns of navigation i.e. direct, pacing, lapping and random. Data were collected from nursing homes for five people with dementia. Ground truth was established manually by observation. Two approaches were used to compare the accuracy of the model. Eight different Machine-learning classifiers were applied in WEKA toolbox. The other approach used was deterministic predefined tree-based algorithm. The comparison made between these two different approaches showed that the deterministic algorithm outperformed on this dataset. The authors proposed method to identify single as well as multiple patterns of wandering within an episode. However, the algorithm has certain shortcomings that it was not able to perform optimally in the case of complicated and varied geometrical setting. Labelling of episode for ground truth was done manually, this is very tiresome job consumes lots of time and energy and it is error prone. The algorithm was not tested on finer location data. All these shortcomings open the window of opportunity to do further study in wandering episode detection. This work aims to propose more refined algorithm that can work in diverse scenarios with maximum accuracy. It is hopeful that some heuristics that can establish the ground truth will ease the burden of labelling episodes. Combining more factors such as temporal and behaviour in identification can make the system more robust.A different approach for wandering detectionIn a series of studies Kearns and colleagues ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2016</Year><RecNum>134</RecNum><DisplayText>[79]</DisplayText><record><rec-number>134</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">134</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Nams, Vilis O</author></authors></contributors><titles><title>Movement Path Tortuosity in Free Ambulation: Relationships to Age and Brain Disease</title><secondary-title>IEEE Journal of Biomedical and Health Informatics</secondary-title></titles><periodical><full-title>IEEE Journal of Biomedical and Health Informatics</full-title></periodical><dates><year>2016</year></dates><urls></urls></record></Cite></EndNote>[79], used Ultra-wideband positioning to study random characteristics in the paths of wandering PWD, ignoring Martino-Saltzman's distinctions among Direct, Lapping, Pacing and Random patterns and focusing instead on the amount of tortuosity in each path generated by each subject over time as they moved about conducting normal daily activities in their home ALF environment. A subject’s path was defined byPEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BUERNPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5BUERNPC9BdXRob3I+PFllYXI+MjAxNzwvWWVhcj48UmVj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ADDIN EN.CITE.DATA [70, 73-75] a period of at least 60 seconds of no movement followed by a movement phase that was followed by another 60 seconds of no movement. Path tortuosity was scaled from 1.0 (perfectly straight path) to 2.0 (Brownian motion) and daily means computed for each subject. Approximately 70% of the ALF residents had been clinically diagnosed with dementia. The investigators found path tortuosity was negatively (Pearson’s r = -0.47) and significantly correlated with scores on a key clinical cognitive measure used to evaluate dementia, the Mini Mental State Exam (MMSE). These results support the notion that the amount of randomness in walking paths can be a useful tool for studying the relationships of movement variability and cognition in older adults. One of the drawbacks of Kearns et al.'s work is that their analytical framework was one-dimensional and collapsed across all of the Martino-Saltzman categories believed by Algase and others to hold significance for the study of dementia. To date there has been very little work examining the fine-grained structure of these patterns.Wandering typology identificationWhile the number of studies related to wandering detection has increased in the recent past, there are still significant gaps in some domains such as identification of particular patterns such as random, looping and pacing in wandering. This creates significant variation in practice associated with the intervention of wandering. It is still debatable whether the benefit outweighs the harmful behaviour associated with wandering. In the past, majority of literature focused on the identification of wandering from the non-wandering patterns. Although these systems were very good at alerting the caregivers as soon as wandering occurred. However, there has been very little work into the fine-grained structure of these patterns.Off the shelves devices for wandering management Plenty of portable, minimally intrusive movement tracking devices such as Opal Sensor (APDM, Inc., Portland, Oregon) ADDIN EN.CITE <EndNote><Cite><Author>APDM</Author><Year>2017</Year><RecNum>198</RecNum><DisplayText>[73]</DisplayText><record><rec-number>198</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">198</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Web Page">12</ref-type><contributors><authors><author>APDM</author></authors></contributors><titles><title>APDM Wearable Technologies</title></titles><volume>2017</volume><number>22/09/2017</number><dates><year>2017</year></dates><urls><related-urls><url></url></related-urls></urls></record></Cite></EndNote>[73], PocketFinder ADDIN EN.CITE <EndNote><Cite><Author>Scalisi</Author><Year>2007</Year><RecNum>195</RecNum><DisplayText>[75]</DisplayText><record><rec-number>195</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">195</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Generic">13</ref-type><contributors><authors><author>Scalisi, Joseph F</author><author>Hecox, Lawrence E</author></authors></contributors><titles><title>Communication system and method including communication channel mobility</title></titles><dates><year>2007</year></dates><publisher>Google Patents</publisher><urls></urls></record></Cite></EndNote>[75], Xsens Motion Tracker ADDIN EN.CITE <EndNote><Cite><Author>Roetenberg</Author><Year>2009</Year><RecNum>167</RecNum><DisplayText>[74]</DisplayText><record><rec-number>167</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">167</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Roetenberg, Daniel</author><author>Luinge, Henk</author><author>Slycke, Per</author></authors></contributors><titles><title>Xsens MVN: full 6DOF human motion tracking using miniature inertial sensors</title><secondary-title>Xsens Motion Technologies BV, Tech. Rep</secondary-title></titles><periodical><full-title>Xsens Motion Technologies BV, Tech. Rep</full-title></periodical><dates><year>2009</year></dates><urls></urls></record></Cite></EndNote>[74], are available that can be used for tracking. Adoption of technology by elderly and healthcare professionalThe application of technology to the study of movement as an indicator of health is receiving significant attention in the health care industry ADDIN EN.CITE <EndNote><Cite><Author>Park</Author><Year>2003</Year><RecNum>516</RecNum><DisplayText>[80]</DisplayText><record><rec-number>516</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516265061">516</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Park, Sungmee</author><author>Jayaraman, Sundaresan</author></authors></contributors><titles><title>Enhancing the quality of life through wearable technology</title><secondary-title>IEEE Engineering in medicine and biology magazine</secondary-title></titles><periodical><full-title>IEEE Engineering in medicine and biology magazine</full-title></periodical><pages>41-48</pages><volume>22</volume><number>3</number><dates><year>2003</year></dates><isbn>0739-5175</isbn><urls></urls></record></Cite></EndNote>[80]. The increasing processing speed, power management and ability to perform a complex computational task of ubiquitous devices, has generated keen interest in their usability. The increased attention is marked by an increased willingness by healthcare professionals to adopt these devices as adjunctive sources of information on their patients ADDIN EN.CITE <EndNote><Cite><Author>Wu</Author><Year>2007</Year><RecNum>517</RecNum><DisplayText>[81]</DisplayText><record><rec-number>517</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516265076">517</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Wu, Jen-Her</author><author>Wang, Shu-Ching</author><author>Lin, Li-Min</author></authors></contributors><titles><title>Mobile computing acceptance factors in the healthcare industry: A structural equation model</title><secondary-title>International journal of medical informatics</secondary-title></titles><periodical><full-title>International journal of medical informatics</full-title></periodical><pages>66-77</pages><volume>76</volume><number>1</number><dates><year>2007</year></dates><isbn>1386-5056</isbn><urls></urls></record></Cite></EndNote>[81]. Patient reluctance to use such devices has also declined as they have become increasingly fashionable. Activity trackers: Fitbit ADDIN EN.CITE <EndNote><Cite><Author>Takacs</Author><Year>2014</Year><RecNum>518</RecNum><DisplayText>[82]</DisplayText><record><rec-number>518</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516265082">518</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Takacs, Judit</author><author>Pollock, Courtney L</author><author>Guenther, Jerrad R</author><author>Bahar, Mohammadreza</author><author>Napier, Christopher</author><author>Hunt, Michael A</author></authors></contributors><titles><title>Validation of the Fitbit One activity monitor device during treadmill walking</title><secondary-title>Journal of Science and Medicine in Sport</secondary-title></titles><periodical><full-title>Journal of Science and Medicine in Sport</full-title></periodical><pages>496-500</pages><volume>17</volume><number>5</number><dates><year>2014</year></dates><isbn>1440-2440</isbn><urls></urls></record></Cite></EndNote>[82], smart-watch, etc. are increasingly becoming an integral part of one’s life and transforming many aspects of clinical practice.by providing physicians with supplementary health information gathered by the patient’s device. Data generated by these devices may play a pivotal role in advancing the development of personalized health care services.With the increasing proportion of elderly presenting health issues associated with later life stages ADDIN EN.CITE <EndNote><Cite><Author>Christensen</Author><Year>2009</Year><RecNum>519</RecNum><DisplayText>[83]</DisplayText><record><rec-number>519</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516265088">519</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Christensen, Kaare</author><author>Doblhammer, Gabriele</author><author>Rau, Roland</author><author>Vaupel, James W</author></authors></contributors><titles><title>Ageing populations: the challenges ahead</title><secondary-title>The lancet</secondary-title></titles><periodical><full-title>The Lancet</full-title></periodical><pages>1196-1208</pages><volume>374</volume><number>9696</number><dates><year>2009</year></dates><isbn>0140-6736</isbn><urls></urls></record></Cite></EndNote>[83], healthcare facilities strain to provide quality health services. A major challenge faced by older adults is dementia, whose symptoms include loss of cognitive functions of a person often accompanied by significant behavioral changes associated with wandering ADDIN EN.CITE <EndNote><Cite><Author>Finkel</Author><Year>1997</Year><RecNum>520</RecNum><DisplayText>[84]</DisplayText><record><rec-number>520</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1516265095">520</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Finkel, Sanford I</author><author>e Silva, Jorge Costa</author><author>Cohen, Gene</author><author>Miller, Sheldon</author><author>Sartorius, Norman</author></authors></contributors><titles><title>Behavioral and psychological signs and symptoms of dementia: a consensus statement on current knowledge and implications for research and treatment</title><secondary-title>International Psychogeriatrics</secondary-title></titles><periodical><full-title>International Psychogeriatrics</full-title></periodical><pages>497-500</pages><volume>8</volume><number>S3</number><dates><year>1997</year></dates><isbn>1741-203X</isbn><urls></urls></record></Cite></EndNote>[84] such as agitation and restlessness; searching or scanning behaviour, hovering or other stationary behavior, and pacing, and an increased potential for unsupervised exiting, or elopement.MotivationThe review of the literature has given us definite insight into the current system for wandering detection and management. The system developed so far has some shortcomings and there are numerous opportunities for improvements. Some of the features and limitation of the current systems are:Lack of ubiquity in the development of wandering detection algorithm: Virtually no attempt has been made by previous researchers to develop wandering detection tool, which can be used ubiquitously in indoor as well as the outdoor environment. Many of the devices have been evaluated in the experimental or special condition and translating these devices into real-world condition is challenging. Even the sensor modalities constrained the use of such tool. The tools were developed to work with only a specific kind of sensor.Large dataset for statistical significance: With very few exceptions almost all the studies have been conducted on very small user groups and for a very small duration of the data collection. This is due to the fact that installation of sensor infrastructure and collection of data is very expensive, and privacy also forms a big hurdle for such data collection. As a result, public dataset comprising of large-scale user trails and experimental results are not yet available.Real data set with correlation with MMSE/RAWS-CV: There have been very few studies which focused on correlation of wandering with cognitive status of person with dementia. Due to the lack of longitudinal dataset it is difficult to identify the statistically significant factor of wandering which correlates with the cognitive state of a person. ?Fine-grained features of wandering and contextual analysis: Although there has been a large number of assistive devices but the technological solution for evaluating various aspects of wandering are relatively fewer and simpler. There have been very few studies which focused on analyzing fine-grained features in wandering trajectory. Most of it focused on differentiating wandering from non-wandering patterns.Management rather than restrictive nature of wandering: For tracking outdoor wandering behaviour, most of the solution focused on restricting wandering rather than management aspects of wandering.Our research has been motivated by following five main objectives:Support for ubiquity in wandering management: We are guided by the principle of development of a ubiquitous system to facilitate large-scale data collection using various sensor modalities (GPS, WIFI, RFID, etc.) applicable to indoor as well as the outdoor environment.A longitudinal study of wandering: There has been no research to date which discusses the longitudinal wandering detection/management methods. This is mainly due to unavailability of a long-term dataset. This is a very promising research filed in order to give an in-depth insight of progression of wandering or the cause of it. This form a major part of our research objective.Correlation study on the wandering with a cognitive test: A statistical data mining approach to unveiling hidden relation between wandering and factors such as cognition opens a big opportunity for the medical science. Factorial analysis of wandering and reliability of such methods forms the objective of our research.Fine-grained feature analysis in wandering: Due to a plethora of sensing devices, a large amount of data collection is possible. These data can capture spatial, temporal as well as ambience features in navigation. Analyzing the data and harnessing environmental factors which can aid in detection and identification of wandering is our objective. It has been identified features such as path tortuosity, amount of randomness, lapping or pacing motion has a significant correlation with cognitive state of a person. Analyzing these features and discovering new one forms the next research objective.Wandering management infrastructure: Our aim is to build the end to end wandering management infrastructure that can be used for the analysis and making informed decision in the diagnosis of dementia.?The aim of this investigation is to advance the science of wandering management by providing clinicians and researchers with new instruments that will lead to a fuller comprehension of its nature and origin. Our aim is to automate the data analytics process for wandering which can work across all the sensor modalities. The creation of an automated tool to extract lapping, pacing and random movement measures from real-time location data simultaneously from multiple individuals will permit a deeper understanding of wandering behavior because of increased sample size and will enable more precisely targeted interventions, since subtle individual differences become more apparent in longitudinal datasets gathered over significant intervals of weeks or months. Presently there is no specialized tool capable of extracting wandering patterns from real-time tracking data.ConclusionThe purpose of this chapter was to review the aspects of wandering. Wandering in the core functional domains most affected by the behavior: mobility; elimination; eating; bathing, dressing, and grooming; communication; and resting. A secondary purpose was to review the technology developed for the wandering detection and suggest clinical strategies specific to persons with dementia who wander. This also builds the roadmap of our current and future research.Classification of Wandering Patterns in Person with DementiaIn this chapter, the author develops a robust algorithm for indoor and outdoor wandering pattern classification. Much of the previous work in this area addressed the measurement of wandering indoors or outdoors and to the best of our knowledge, there has not been a uni?ed algorithm proposed which can deal with both scenarios. The author presents a novel grid-based layout representation strategy to identify the patterns, which is applicable to both indoor and outdoor scenarios. The algorithm is sufficiently robust to identify interleaving and hybrid patterns and performed very well on a real-world sample. The author begins the chapter with the introduction and then reviews some of the important work done in this field. This is followed by preliminaries which are important to the understanding of the proposed algorithm. The author finally presents the algorithm and the framework, which is an end-to-end implementation of the algorithm from a system point of view. At the end of this chapter, the author also shows the feasibility of such system with a real dataset result. IntroductionResearch in this field points out that people suffering from dementia can show the wandering behavior at a certain stage of their illness. It has been found that around 6 out of 10 people with dementia tend to wander. Wandering is a significant behavioral problem associated with dementia patients. It is a serious concern for family and professional care providers as well as for scientists, clinicians, and policy makers, because the behavior is associated with some of the gravest adverse outcomes in dementia care (e.g., accidents, falls, getting lost and even death). Current research in wandering is focused on one of these four concerns ADDIN EN.CITE <EndNote><Cite><Author>Nelson</Author><Year>2007</Year><RecNum>121</RecNum><DisplayText>[85]</DisplayText><record><rec-number>121</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">121</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Book">6</ref-type><contributors><authors><author>Nelson, Audrey</author><author>Algase, Donna L</author></authors></contributors><titles><title>Evidence-based protocols for managing wandering behaviors</title></titles><dates><year>2007</year></dates><publisher>Springer Publishing Company</publisher><isbn>0826163653</isbn><urls></urls></record></Cite></EndNote>[85] as shown in REF _Ref504397851 \h \* MERGEFORMAT Figure 3.1. Figure 3.1: Current field of research in WanderingAspects of this chapter are based on these journal articles, “Mapping the Maze of Terms and Definitions in Dementia-Related Wandering,” ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2007</Year><RecNum>120</RecNum><DisplayText>[86]</DisplayText><record><rec-number>120</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">120</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, DL</author><author>Moore, D Helen</author><author>Vandeweerd, Carla</author><author>Gavin-Dreschnack, DJ</author></authors></contributors><titles><title>Mapping the maze of terms and definitions in dementia-related wandering</title><secondary-title>Aging & mental health</secondary-title></titles><periodical><full-title>Aging Ment Health</full-title><abbr-1>Aging & mental health</abbr-1></periodical><pages>686-698</pages><volume>11</volume><number>6</number><dates><year>2007</year></dates><isbn>1360-7863</isbn><urls></urls></record></Cite></EndNote>[86] and “Travel behavior of nursing home residents perceived as wanderers and non-wanderers” ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>97</RecNum><DisplayText>[7]</DisplayText><record><rec-number>97</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">97</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7]. The article ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2007</Year><RecNum>120</RecNum><DisplayText>[86]</DisplayText><record><rec-number>120</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">120</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, DL</author><author>Moore, D Helen</author><author>Vandeweerd, Carla</author><author>Gavin-Dreschnack, DJ</author></authors></contributors><titles><title>Mapping the maze of terms and definitions in dementia-related wandering</title><secondary-title>Aging & mental health</secondary-title></titles><periodical><full-title>Aging Ment Health</full-title><abbr-1>Aging & mental health</abbr-1></periodical><pages>686-698</pages><volume>11</volume><number>6</number><dates><year>2007</year></dates><isbn>1360-7863</isbn><urls></urls></record></Cite></EndNote>[86] aims to standardize the language in wandering science by giving an operational definition of dementia-related wandering, which can aid in clinical recognition, to promote research precision and validity. In ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>97</RecNum><DisplayText>[7]</DisplayText><record><rec-number>97</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">97</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7], Martino-Saltzman classified independent travel into four patterns ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>97</RecNum><DisplayText>[7]</DisplayText><record><rec-number>97</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">97</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7] direct, random, pacing and lapping as illustrated in REF _Ref417264051 \h \* MERGEFORMAT Figure 3.2. Figure 3.2: Patterns of ambulation (Martino-Salzman, 1991)Out of these four patterns, random, pacing and lapping are considered as wandering. Direct pattern is not considered as a sign of wandering because this is an efficient way of travelling. Our research to differentiate wandering from non-wandering is based on this classification method. Our aim is to automatically classify the movement traced by patients into one of these navigation patterns. In a study conducted by Algase ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2001</Year><RecNum>150</RecNum><DisplayText>[8]</DisplayText><record><rec-number>150</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">150</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author><author>Beattie, Elizabeth RA</author><author>Therrien, Barbara</author></authors></contributors><titles><title>Impact of cognitive impairment on wandering behavior</title><secondary-title>Western Journal of Nursing Research</secondary-title></titles><periodical><full-title>Western Journal of Nursing Research</full-title></periodical><pages>283-295</pages><volume>23</volume><number>3</number><dates><year>2001</year></dates><isbn>0193-9459</isbn><urls></urls></record></Cite></EndNote>[8] on 25 demented residents from two long-term care settings, the team explored cognitive impairment as a predictor of wandering rhythm (cycle duration, frequency and structure) and pattern (random, lapping and pacing). Algase classified wandering into 7 subscales as described in section REF _Ref522116186 \r \h 2.4.2, the three subscales were significantly correlated to the hourly rate and the percent and duration of wandering episodes (in proportion to all walking) for overall wandering (random, lapping, and pacing together) and for random pattern wandering alone. But persistent walking and eloping behaviour subscales were correlated with lapping and pacing pattern wandering. The study has shown that the pattern of relationships among pacing and other dimensions of wandering (persistent walking, spatial disorientation, eloping behaviour, and repetitive walking) differs from that of other wandering, that is, random and lapping. This gives us the motivation to study each of these patterns separately to unearth any relationship with cognitive state of person with dementia. To the best of our knowledge, there has been no reliable and accurate method which can automatically detect and classify the wandering pattern of a dementia patient. This activity is generally performed by a trained professional who observes the patient movement on camera and label the pattern manually into the system. This method is very inefficient and error prone. Besides, it also requires plenty of manpower to observe the patterns. Some of the research done for automatic detection of wandering ADDIN EN.CITE <EndNote><Cite><Author>Vuong</Author><Year>2014</Year><RecNum>99</RecNum><DisplayText>[77, 78]</DisplayText><record><rec-number>99</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">99</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Vuong, NK</author><author>Chan, S</author><author>Lau, CT</author></authors></contributors><titles><title>Automated detection of wandering patterns in people with dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>127-147</pages><volume>12</volume><number>3</number><dates><year>2014</year></dates><isbn>1569-111X</isbn><urls></urls></record></Cite><Cite><Author>Vuong</Author><Year>2013</Year><RecNum>122</RecNum><record><rec-number>122</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">122</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Vuong, NK</author><author>Goh, SGA</author><author>Chan, S</author><author>Lau, CT</author></authors></contributors><titles><title>A mobile-health application to detect wandering patterns of elderly people in home environment</title><secondary-title>Engineering in Medicine and Biology Society (EMBC), 2013 35th Annual International Conference of the IEEE</secondary-title></titles><pages>6748-6751</pages><dates><year>2013</year></dates><publisher>IEEE</publisher><isbn>1557-170X</isbn><urls></urls></record></Cite></EndNote>[77, 78] has the limitation that it is constrained for particular layout of the facility and cannot be applied to the fine-grained structure, for example, the layout used by Kearns in “Wireless telesurveillance system for detecting dementia” ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2011</Year><RecNum>95</RecNum><DisplayText>[87]</DisplayText><record><rec-number>95</rec-number><foreign-keys><key app="EN" db-id="detf9ptwczae9se0exnpdrfqv2fwww52e9wv" timestamp="1428945088">95</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Nams, Vilis O</author><author>Craighead, Jeffrey D</author></authors></contributors><titles><title>Wireless telesurveillance system for detecting dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>90-102. doi: 10.4017/gt. 2011.10. 2.004. 00</pages><volume>10</volume><number>2</number><dates><year>2011</year></dates><urls></urls></record></Cite></EndNote>[87]. Wandering behaviour can also occur both indoors and outdoors. When it occurs outdoors, dire consequences such as getting lost, falling or elopement may result. Attempts have been made to identify these patterns taking into consideration indoor or outdoor behavior ADDIN EN.CITE <EndNote><Cite><Author>Lin</Author><Year>2012</Year><RecNum>131</RecNum><DisplayText>[19, 78]</DisplayText><record><rec-number>131</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">131</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Lin, Qiang</author><author>Zhang, Daqing</author><author>Huang, Xiaodi</author><author>Ni, Hongbo</author><author>Zhou, Xingshe</author></authors></contributors><titles><title>Detecting wandering behavior based on GPS traces for elders with dementia</title><secondary-title>Control Automation Robotics & Vision (ICARCV), 2012 12th International Conference on</secondary-title></titles><pages>672-677</pages><dates><year>2012</year></dates><publisher>IEEE</publisher><isbn>146731871X</isbn><urls></urls></record></Cite><Cite><Author>Vuong</Author><Year>2014</Year><RecNum>130</RecNum><record><rec-number>130</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">130</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Vuong, NK</author><author>Chan, S</author><author>Lau, CT</author></authors></contributors><titles><title>Automated detection of wandering patterns in people with dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>127-147</pages><volume>12</volume><number>3</number><dates><year>2014</year></dates><isbn>1569-111X</isbn><urls></urls></record></Cite></EndNote>[19, 78] but not both. This poses a serious problem in system design for the practical application of wandering detection. The author propose a grid-based approach for detecting wandering which can efficiently handle both domains and facilitates data analysis of these patterns.Specifically, contributions in this chapter can be summarized as follows:The author presents a novel algorithm providing an end-to-end solution for wandering management. The algorithm is universally applicable for both indoor and outdoor navigation and is agnostic with respect to the source of the location data.The author classifies each navigational episode into one of the four patterns of navigation using Martino-Saltzman classification and calculate statistical features of the episode.The author also illustrates the modeling of paths using grid-based layout representation. We then apply our algorithm in the next chapter on a real-world dataset and report the performance of the algorithm in terms of precision, recall and accuracy.The author finally proposes the framework for the complete end-to-end management of the wandering behaviour.Related WorkWith the proliferation of new sensing, computing and communication devices, inexpensive technological solutions have been developed to tackle the problems associated with wandering. iWander ADDIN EN.CITE <EndNote><Cite><Author>Sposaro</Author><Year>2010</Year><RecNum>132</RecNum><DisplayText>[88]</DisplayText><record><rec-number>132</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">132</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Sposaro, Frank</author><author>Danielson, Justin</author><author>Tyson, Gary</author></authors></contributors><titles><title>iWander: An Android application for dementia patients</title><secondary-title>Engineering in Medicine and Biology Society (EMBC), 2010 Annual International Conference of the IEEE</secondary-title></titles><pages>3875-3878</pages><dates><year>2010</year></dates><publisher>IEEE</publisher><isbn>1424441234</isbn><urls></urls></record></Cite></EndNote>[88], is an Android based application which predicts wandering from the contextual information collected from mobile sensors and sends noti?cation to the caretakers. It can also assist persons with dementia (PWD) in the process of navigating their local environment.Focusing on outdoor wandering, Lin et al. ADDIN EN.CITE <EndNote><Cite><Author>Lin</Author><Year>2012</Year><RecNum>131</RecNum><DisplayText>[19]</DisplayText><record><rec-number>131</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">131</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Conference Proceedings">10</ref-type><contributors><authors><author>Lin, Qiang</author><author>Zhang, Daqing</author><author>Huang, Xiaodi</author><author>Ni, Hongbo</author><author>Zhou, Xingshe</author></authors></contributors><titles><title>Detecting wandering behavior based on GPS traces for elders with dementia</title><secondary-title>Control Automation Robotics & Vision (ICARCV), 2012 12th International Conference on</secondary-title></titles><pages>672-677</pages><dates><year>2012</year></dates><publisher>IEEE</publisher><isbn>146731871X</isbn><urls></urls></record></Cite></EndNote>[19] have proposed a data driven model for the detection of lapping and pacing based on the user's GPS traces, where the angular sum of the turning points has been used to identify lapping and pacing motions. N.K. Vuong et al. ADDIN EN.CITE <EndNote><Cite><Author>Vuong</Author><Year>2014</Year><RecNum>130</RecNum><DisplayText>[78]</DisplayText><record><rec-number>130</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">130</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Vuong, NK</author><author>Chan, S</author><author>Lau, CT</author></authors></contributors><titles><title>Automated detection of wandering patterns in people with dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>127-147</pages><volume>12</volume><number>3</number><dates><year>2014</year></dates><isbn>1569-111X</isbn><urls></urls></record></Cite></EndNote>[78] carried out a systematic study of the automatic classi?cation of wandering patterns. A set of machine learning and deterministic algorithms were used to compare the accuracy of the methods used. It was found that tree-based deterministic method was most suitable for real-time application.Most of the literature has focused on distinguishing wandering from the direct pattern. However, some researchers have explored the characteristics of wandering based on the geographical pattern of the behavior. In a series of studies Kearns and colleagues ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2016</Year><RecNum>134</RecNum><DisplayText>[79]</DisplayText><record><rec-number>134</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">134</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Nams, Vilis O</author></authors></contributors><titles><title>Movement Path Tortuosity in Free Ambulation: Relationships to Age and Brain Disease</title><secondary-title>IEEE Journal of Biomedical and Health Informatics</secondary-title></titles><periodical><full-title>IEEE Journal of Biomedical and Health Informatics</full-title></periodical><dates><year>2016</year></dates><urls></urls></record></Cite></EndNote>[79], used Ultra-wideband positioning to study random characteristics in the paths of wandering PWD, ignoring Martino-Saltzman's distinctions among Direct, Lapping, Pacing and Random patterns and focusing instead on the amount of tortuosity in each path generated by each subject over time as they moved about conducting normal daily activities in their home ALF environment. PreliminariesIn this section, the author discusses some of the key concepts which have been considered while designing the algorithm. Physical layout RepresentationOne of the important consideration made to design the algorithm has been the representation of physical layout as a set of grids. In this technique, the world is represented as an array of blocked and unblocked cells. Grid representation allows quick access to the layout and supports navigational tasks such as path planning. The approach confers the following advantages: Localization is easy, simply requiring the coordinates to be divided by the grid resolution sizeObstacles can be easily representedPath planning is possible using heuristics such as the A* algorithmPath efficiency/estimation and the representation of the path in the form of graph/tensor is possibleNoise reduction properties are built-in Grid resolution can be varied to efficiently represent the pathDisadvantages of the approach include introduction of potential discretization errors, and high computational memory requirements.With respect to our problem statement, grid layout provides the optimal way for physical landscape representation for indoor as well as outdoor navigation. Grid has also been traditionally used in a number of cases for path planning in robot navigation ADDIN EN.CITE <EndNote><Cite><Author>Hachour</Author><Year>2008</Year><RecNum>135</RecNum><DisplayText>[89]</DisplayText><record><rec-number>135</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">135</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Hachour, O</author></authors></contributors><titles><title>Path planning of Autonomous Mobile robot</title><secondary-title>International journal of systems applications, engineering & development</secondary-title></titles><periodical><full-title>International journal of systems applications, engineering & development</full-title></periodical><pages>178-190</pages><volume>2</volume><number>4</number><dates><year>2008</year></dates><urls></urls></record></Cite></EndNote>[89]. To the best of our knowledge, this is the ?rst time it has been used in wandering pattern representation. Wandering typology in a grid map representationThe author defines the typology with respect to grid layout representation. Traditionally the typology has been verbally defined ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>129</RecNum><DisplayText>[7]</DisplayText><record><rec-number>129</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">129</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7] for ease of understanding, as usually human perception was the only means for labeling the observed typologies. However, with the torrent of navigation data arising from tracking PWD it has virtually become impossible to do this manually, and the process has been prone to human error. This has necessitated the requirement to redefine the typology such that mathematical modeling can be utilized to automatically label patterns and preserve the original definition statement.A navigation pattern can be represented as shown in REF _Ref502327974 \h \* MERGEFORMAT Figure 3.4. Here, the Direct pattern can be defined as non-intersecting grid path from source to destination up to a certain level of optimality. The Random path can be non-intersecting as well as intersecting with a very low level of optimality. Lapping and Pacing are looping patterns with at least two consecutive loops, roughly overlapping each loop. The distinction between Lapping and Pacing can be made according to the area enclosed by the loop. Even if a pattern has a loop but does not conform to the requirement of the number of loops or overlapping conditions, it will be considered as a Random pattern. In our grid layout representation, navigation from one grid to the adjacent grid is allowed only through edge movement, no diagonal movement is permitted as shown in REF _Ref502329072 \h \* MERGEFORMAT Figure 3.3. This minimizes much of the representational and computational overhead. Some of the trajectory information can be lost by this constraint but by carefully selecting the grid size this error can be minimized. As a heuristic, we select the grid size given by:Grid_Size=maxstride_length, Sensor_accuracyHere, stride_length is the roughly 75cm for normal walking motion. Sensor_accuracy for indoor is in the range of 20cm whereas for GPS it is 2m. For mapping indoor navigation Grid_Size = 0.75m was used, whereas for outdoor navigation the Grid_Size = 2m was appropriate. By using such heuristic, we minimized the discretization error and at the same time also automatically removed the noise from data. Figure STYLEREF 1 \s 3. SEQ Figure \* ARABIC \s 1 3: Direction of navigation from gridFigure 3.4: Navigational Pattern in grid world representationThe Proposed AlgorithmIn this section the author discusses the proposed algorithm. There are three stages as shown in REF _Ref502328413 \h \* MERGEFORMAT Figure 3.5. The first stage is data pre-processing, followed by segmentation, and finally classification of the navigational pattern according to Martino-Saltzman proposed method. Each of these steps is discussed in detail below: Figure 3.5: The Workflow FrameworkData Pre-processing The first stage processes data in two steps - data cleaning and trajectory compression. Data Cleaning: Spatiotemporal data may contain inaccuracies due to sensor noise and other factors. Sensor data is first passed through a heuristic-based outlier detector. We also found that when we plotted navigation paths in a time-series, some points deviated spatially from the normal path by a very large margin. The heuristic used to identify these points was based on mean speed of the preceding points within the fixed window length. A window size of 5 allowed reliable identification of these outliers. If the speed exceeded the threshold value (four times the average speed in the window) it was considered a potential outlier. The outliers identified using this method ranged from 1% - 4% of total recorded points. Specifically, we consider following class of data for pre-processing task:Inconsistent Data: Inconsistencies in sensor data were found mostly at the corner of the facilities. This may be due to the presence of an iron beams in construction material of the facility. These inconsistent readings were identified by observing change in navigation speed. Speeds were found to be exceptionally high in these cases. Therefore, these points were removed from the reading.Redundant Data: Ubisense devices registered the navigation coordinates in the order of cm. To plot the navigation pattern of a person, not all the points were required. Stepwise analysis was done to find the most appropriate points for consideration. Step-distance, parameter defined the minimum distance between two consecutive points that can be considered for episode plotting. It was selected in such a way that actual movement pattern was not distorted after removing some of the redundant points. From the analysis, it was found that most appropriate size to keep for step-distance was around 0.5m, which is about the same size as a single step taken in normal movement. This is also consistent with the logic that movement pattern will not be distorted if we remove all the points in between the single step. We cannot also keep the value of step-distance very large. In that case, it will smooth out some of the patterns in a movement.Outliers Data: It was also observed that some of the data points were outside the facility dimension and beyond the wall. These points were identified as outliers and removed. After applying all the data cleaning procedures, outline of the facility was identified, and movement episodes were plotted in Matlab. The coordinate points were scaled to fit the outline of the facility. The episodes simulated in Matlab were used to identify the movement patterns, which were further used to establish the ground truth to compare the accuracy of developed model. Trajectory Compression: Due to its very high sampling rate, the Ubisense system generated a large amount of raw data ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2011</Year><RecNum>8837</RecNum><DisplayText>[90]</DisplayText><record><rec-number>8837</rec-number><foreign-keys><key app="EN" db-id="avf5v0ptmz900ne2xem5999f2tzfas2pdzsd" timestamp="1487280418">8837</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, W.</author><author>Fozard, J. L.</author><author>Nams, V. O.</author><author>Craighead, J. D.</author></authors></contributors><titles><title>Wireless telesurveillance system for detecting dementia</title><secondary-title>Gerontechnology</secondary-title><short-title>Wireless telesurveillance system for detecting dementia.</short-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>90-102</pages><volume>10</volume><number>2</number><section>90</section><dates><year>2011</year><pub-dates><date>2011</date></pub-dates></dates><urls></urls></record></Cite></EndNote>[90]. Using these data verbatim would require significant computing power and energy, so we compressed the trajectory using mean-clustering. Data collected from the Ubisense system can be represented in a sequence as p={p1,p2,…,pI,…,pn}. Each point pi∈p contains spatial and temporal information for the position. We define the radius parameter of the cluster as dthres, this value can be fine-tuned for the specific case. In our case dthres=0.25 was found to be optimal for preserving the trajectory information. The Distance function is defined as di,j=distancepi,pj, which is Euclidian distance of point pj from pi and speed is defined as si,j=speedpi,pj, which is the travel speed from point i to j. The new set of points after initial processing is represented as burst Bi. The complete algorithm is shown in REF _Ref504401099 \h \* MERGEFORMAT Algorithm 3.1. Algorithm 3.1: Trajectory Compression Algorithm REF _Ref490558274 \h \* MERGEFORMAT Figure 3.6 represents one such case of high-sampled data. We processed the data in sequential order described in algorithm 1. Initial Point pi is directly coded as Bi, we calculate d1,2, here d1,2≥dthres so D2=p2, this is the case with point p2 and p3, but d4,5<dthres, so we calculate the average of these points as pcen=p4+p5/2. If it is greater than dthres, we record the new point as first of the next burst, or else we keep repeating the process until we obtain a point whose distance from the previous point is greater than r. Here point p4 to p7 falls in one cluster with cluster center Pcen as depicted by the solid green ball. The number of data points is reduced by approximately 60% - 70%, resulting in better processing performance without significant loss of path information. Error of trajectory compression is calculated using the perpendicular Euclidean distance ADDIN EN.CITE <EndNote><Cite><Author>Zheng</Author><Year>2015</Year><RecNum>182</RecNum><DisplayText>[91]</DisplayText><record><rec-number>182</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">182</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Zheng, Yu</author></authors></contributors><titles><title>Trajectory data mining: an overview</title><secondary-title>ACM Transactions on Intelligent Systems and Technology (TIST)</secondary-title></titles><periodical><full-title>ACM Transactions on Intelligent Systems and Technology (TIST)</full-title></periodical><pages>29</pages><volume>6</volume><number>3</number><dates><year>2015</year></dates><isbn>2157-6904</isbn><urls></urls></record></Cite></EndNote>[91] method which sums length of the segment perpendicular to the actual trajectory vs. compressed trajectory. This index is always maintained such that compression error does not deviate significantly from a specified threshold. Figure 3.6: Example of trajectory compressionSegmentationNavigation path consists of an alternating locomotion and non-locomotion phases as shown in REF _Ref502419053 \h \* MERGEFORMAT Figure 3.7.a. If there is no motion for more than 60 seconds, it is considered as a non-locomotion phase. Each locomotion phase constitutes an episode in navigation as shown in REF _Ref502419053 \h \* MERGEFORMAT Figure 3.7.b. It consists of a burst of relocations which contain spatial and temporal information of an episode. Bursts are generally non-uniformly distributed in spatial and temporal co-ordinates. To simplify the navigation path, we compute distance and angle between two consecutive bursts as shown in REF _Ref502419053 \h \* MERGEFORMAT Figure 3.7.c. and divide them into equal step-lengths as shown in REF _Ref502419053 \h \* MERGEFORMAT Figure 3.7.d. A step-length of 0.4m was found to give the best approximation of actual step-length for casual walking motion. In one step, navigation is only allowed to only one adjacent grid. This simpli?es the statistical and features calculation step. An episode can be represented as: where, : No. of episodes, and : No. of bursts in an episode. A burst can be represented as: : burst for the navigation pathAfter discretization, an episode is divided into steps that can be represented as a vector of length step-length:dwhere, , m is total steps and d is a unit vector in step direction.In the grid layout we represent an episode as: So, a unique episode in the grid layout system is represented by:where, is start co-ordinate for an episode.Figure 3.7: Step Discretization method in episode plottingStay-point detection:We partition navigation into ambulatory and sedentary phases and apply our algorithm to the ambulatory phase. One challenge is reliably identifying the sedentary phase given sensor noise, which can falsely indicate movement. Empirical examination showed that noisy data usually vary significantly from neighboring points in their apparent speed and by the number of sharp turns between adjacent points. These considerations substantially reduced the noise level in the data.Segmentation: We defined a sedentary phase as a segment > 60s with no movement registered by the sensor. Smoothing is done over 10-points using the mean method to reduce small fluctuations in sensor readings and minimize false movements. A threshold value of 0.2m/s optimally partitions the sensor data into locomotion and sedentary phases. ClassificationFeatures Calculation: A set of features was calculated for each episode. These features were used to identify the travelling pattern of patients. The features along with its explanation are listed in the REF _Ref504414221 \h \* MERGEFORMAT Table 3.1.Table 3.1: Features of the episodeFeaturesExplanationx_sumSum of steps taken in x-direction (with sign)y_sumSum of steps taken in y-direction (with sign)xy_sumExplained belowx_abs_sumSum of steps taken in x-direction (without sign)y_abs_sumSum of steps taken in y-direction (without sign)xy_abs_sumSum of steps taken in x-direction and y-direction (without sign)angle_sumSum of angles turnedangle_abs_sumSum of absolute angle turnedno_of_roundsNo of round takenExplanation of the features: xy_sumFor e.g. feature xy_sum is calculated by the formula: xy_sum i,j =abs[Sum(xi to xj)] + abs[Sum(yi to yj)]In this, each pair of steps has been created as a lookup table in the symmetric matrix of dimension NxN. Here N is the total no of steps in an episode.Episode Classification: From the analysis of an actual episode, it has been found that each episode can be very complex and can be split into simpler looping/non-looping segment. Here a looping segment is defined as the largest continuous segment which intersects with itself whereas a non-looping segment does not intersect. Algorithm 3.2: Steps to find loop in a pathFirst, we process non-looping segment, which can be either random or direct. To find the optimal path from source to destination we use A* algorithm given by Equation (1). Here, Layout represents the physical layout in grid world representation. Actual path length is scalar sum of steps as calculated from Equation (2) and, Path efficiency η is calculated from Equation (3). This is an important parameter which is used to arrive at the optimal labelling of the segment. It has been found in our case that η = 0.6 gives the most logical and accurate result for the segment. (1) (2) (3)Next, REF _Ref504412791 \h \* MERGEFORMAT Algorithm 3.2 detects any loop in an episode. Each loop is processed separately to ?nd wandering patterns i.e. lapping, pacing or random. Lapping and pacing are repetitive in nature with at-least two loops. Distinction between lapping and pacing is done by finding the area enclosed within the loop. If the area is less than minimum area possible with the given segment length, it is considered as pacing, or else it is classed as lapping. Finally, consolidation is done for the last pattern. We also calculate time taken and distance travelled for each pattern type.Ambulation features are calculated and passed to the episode identification algorithm to classify it as direct travel, random travel, lapping or pacing REF _Ref490562597 \h \* MERGEFORMAT Figure 3.8, shows the complete flowchart for episode classification. Figure 3.8: Flowchart for episode consolidationA sample episode representation in grid worldWe show a sample path followed by a subject in the physical layout. Grid Representation of Layout: We have divided the layout into equally sized grid as shown in REF _Ref417264269 \h \* MERGEFORMAT Figure 3.9. Each grid dimension is equal to a pre-defined size called step_size. The same distance has been used to remove redundant points from data sets. This was also found to be the optimal size. Here it has been taken as 0.5m for all the further implementation and processing.Figure 3.9: Layout divided into equally sized gridEach episode is plotted as movement from one grid to adjacent grid. Movement is constrained only to +X, -X, +Y or –Y direction. Diagonal movement has been excluded for simplicity. Angle for each step taken is either 0(+X), 90(+Y), 180(-X) or 270(-Y)Each movement in an episode has been approximated in accordance with the above-mentioned rule. A sample episode movement looks like as shown in REF _Ref417264344 \h \* MERGEFORMAT Figure 3.10. 1800860149415512382501866900Loop00Loop35509061426329Smoothing00Smoothing39282621101039Figure 3.10: Sample episode plot in MatlabThis procedure smoothed out path from start to end, and at the same time it is able to capture the loop in a path. Size of step_size will define the smoothing factor. It can be noticed that, ifStep_size is too large; it will smooth out the actual looping episode.Step_size is too small; due to noise in sensor reading, it will produce the false loop in an episode.So, there must be a balance between these two competing factors. By intuition, step_size must be greater than offset produced in sensor reading due to noise and must be less than the size of the human step so that it can capture all the relevant motion. Each movement from one grid to adjacent grid is captured in step identified by following notation shown in REF _Ref504414339 \h \* MERGEFORMAT Table 3.2:Table 3.2: Coding of the navigation directionXYAngleRemark100Moving in +ve X direction0190Moving in +ve Y direction-10180Moving in -ve X direction0-1270Moving in -ve Y directionEach episode is represented by a step sequence. This makes easy to compute features of each episode. Discretized sample episode will look as shown in REF _Ref417264344 \h \* MERGEFORMAT Figure 3.10.Framework (RT-WMAT): Real-time Wandering Management & Analytics ToolIn this section, the author discusses the framework proposed for the end-to-end implementation of the algorithm discussed in the previous section. It starts with a brief introduction and list down the drawbacks of the current system, and after carefully reviewing previous methods the author lists down four improvements which have been incorporated in our proposed framework. Detection and management of wandering forms the active part of research in the recent time. A system to detect wandering from the GPS traces of person in real time is quite accurate and can be ubiquitous [8]. However, it does not provide any contextual information about wandering e.g. place and time of the wandering. This information can give extra insight into the nature of wandering and proper management techniques can be implemented. Navigation traces also have features such as intensity of randomness, average path length and total time spent in wandering, etc. Randomness intensity can be captured as the measure of tortuosity (Fractal-D), which is a significant predictor of dementia [6]. For example, iWander [12] is an Android based application which predicts wandering from the data collected from a mobile sensor and sends the notification to the caretakers. Here, there is no provision that these valuable data can be stored for future use and data mining techniques can be applied to find some patterns or trend. There is also no provision for report generation of navigation pattern for the dementia patients. These applications for wandering detection collect spatio-temporal and contextual information from the environment but do not utilize the data to its full potential. The previous researchers have used these data only for the simple pattern recognition task and somehow ignored the long-time benefit which can arise by properly managing, storing and analyzing these data. After reviewing several wandering management tools, we found the four most important improvement which can be applied to the existing solutions: Improvement I1: Design of a data management tool for wandering.Improvement I2: Capturing contextual and temporal information about the wandering.Improvement I3: Development of a tool to generate the exhaustive list of features of wandering for the task of data mining and analysis.Improvement I4: Development of reporting tool which can generate a long-term report on wandering and generate actionable interpretation from the data. With respect to our first proposed improvement I1, we have found that in the last two decades, significant progress has been made in pervasive computing technologies which have made possible the fast, easy and efficient deployment of sensor devices in the environment. Recently, there has been a flood of available sensor data due to the widespread use of these devices. However, management of these critical data is still not explored. Apart from the navigational data each patient also has clinical report such as MMSE or RAWS-CV, which are updated periodically based on the suggestion of the physician. Combining these two types of data to give us insight into the dementia progression is the goal for our work.Sensor devices can also capture the contextual information such as spatial, temporal or vital signs relating to the specific situation. The improvement I2 suggests that some of the information such as place, time and intensity of wandering can play very important role in the design of the environment which can restrict wandering. Researchers have found that certain geographical layout can intensify wandering ADDIN CSL_CITATION { "citationItems" : [ { "id" : "ITEM-1", "itemData" : { "ISSN" : "0265-8135", "author" : [ { "dropping-particle" : "", "family" : "Mitchell", "given" : "Lynne", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Burton", "given" : "Elizabeth", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Raman", "given" : "Shibu", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Blackman", "given" : "Tim", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Jenks", "given" : "Mike", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Williams", "given" : "Katie", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" } ], "container-title" : "Environment and Planning B: Planning and Design", "id" : "ITEM-1", "issue" : "4", "issued" : { "date-parts" : [ [ "2003" ] ] }, "page" : "605-632", "publisher" : "SAGE Publications", "title" : "Making the outside world dementia-friendly: design issues and considerations", "type" : "article-journal", "volume" : "30" }, "uris" : [ "" ] } ], "mendeley" : { "formattedCitation" : "[10]", "plainTextFormattedCitation" : "[10]", "previouslyFormattedCitation" : "[10]" }, "properties" : { "noteIndex" : 0 }, "schema" : "" }[10]. If the certain area in ALF shows abnormal wandering episodes, sign card can be used to help patients to prevent wandering. For certain patients, wandering episodes intensify at a particular time of day. This information is also critical in the diagnosis of dementia.The improvement I3 and I4, suggest improving the system from the data analytics perspective. We could identify some features in the navigation which correlates significantly with the dementia progression. Calculation of these features over a period and then analyzing the trends can be very insightful in deciding the course of medication for the physician. This will also help in identification of abnormal situation which might not be possible without the aid of proper analytics tool. We also propose to create a reporting tool for wandering which the physician or caretakers can generate on the fly. Finally, a toolbox which can be coupled to location tracking data to facilitate analysis will be of great help in hastening research in this area.In our preliminary study, we have collected indoor/outdoor navigational data of real patients and classified them into navigation patterns as suggested by Martino-Saltzman ADDIN CSL_CITATION { "citationItems" : [ { "id" : "ITEM-1", "itemData" : { "DOI" : "10.1093/geront/31.5.666", "ISBN" : "0016-9013", "ISSN" : "0016-9013", "PMID" : "1778493", "abstract" : "A video-based observational methodology was used to assess the travel behaviors of 40 nursing home residents, 24 of whom were identified by nursing staff as wanderers. Travel was monitored continuously for 30 days, resulting in the recording of over 5,000 unassisted travel events. Four basic travel patterns were observed: direct travel (86.8%), lapping (11.6%), random travel (.9%), and pacing (.7%). Travel efficiency (percentage of direct travel) was significantly related to cognitive status (r = .56), with inefficient travel most prevalent in severely demented participants.", "author" : [ { "dropping-particle" : "", "family" : "Martino-Saltzman", "given" : "D", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Blasch", "given" : "B B", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Morris", "given" : "R D", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "McNeal", "given" : "L W", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" } ], "container-title" : "The Gerontologist", "id" : "ITEM-1", "issue" : "5", "issued" : { "date-parts" : [ [ "1991", "10" ] ] }, "page" : "666-672", "title" : "Travel behavior of nursing home residents perceived as wanderers and nonwanderers.", "type" : "article-journal", "volume" : "31" }, "uris" : [ "" ] } ], "mendeley" : { "formattedCitation" : "[9]", "plainTextFormattedCitation" : "[9]", "previouslyFormattedCitation" : "[9]" }, "properties" : { "noteIndex" : 0 }, "schema" : "" }[9], this classification is based on the geographical region and is the prevailing strategy used by researchers studying wandering. We have used a novel grid-based approach as discussed previously in this chapter. For the latter part of our study, we will be analyzing the navigational patterns of Persons with Dementia (PwD) over a longer duration of time, e.g. 1 year, to isolate trends in the features associated with wandering (e.g. path-efficiency, Fractal-Dimension of the path). If significant differences in the trends between PwD and non-PwD patient groups are found, then these features can be monitored longitudinally to gain a better understanding of the progression of dementia. Our preliminary analysis of trends in the statistical features of navigation for PwD suggests the existence of such differentiation.Goal and Scope of the FrameworkOur goal is to develop a framework for the real-time wandering management and analytic tool for dementia patient (RT-WMAT). This can alleviate the pain in elderly care. It aims to help doctor, caregivers, and family member by providing a tool to understand the nature of wandering. We also aim to make it easy for doctors to study the change in patterns of wandering by trend analysis over a longer duration. This information can be very useful in deciding the course of medication in future and can add an extra dimension to the care of dementia patients. We will propose the key design components of the framework.System ArchitectureTo achieve the goal, we have implemented the solution as a server-client distributed architecture. The server acts as the providers of resource and services and a client is the requester of these services. The architecture allows the multiple clients to connect simultaneously. It is highly scalable, and role-based security in terms of authentication and authorization can be easily implemented. It also provides better data sharing opportunity. We propose to develop an end-to-end system for all the operations associated with wandering management. REF _Ref502660297 \h \* MERGEFORMAT Figure 3.11 shows the overall system design. There are three components of the system.Figure 3.11: A Real-Time Wandering Management and Analytics Tool (RT-WMAT)Sensor devices provide the way to collect spatio-temporal and contextual data from the environment. The centralized server acts as a connecting point for sensor devices and clients. The client can make request for services to the server.Sensors (Spatiotemporal data and contextual information):Sensor technologies used in this field are basically RFID, UWB, Inertial sensor, etc. These sensors can be either embedded in the environment or can be carried by the individual in the form of smartwatch or smartphone. Sensor devices used for the indoor and outdoor devices are different. Outdoor logging device such as GPS does not act well in the indoor environment. For outdoor scenarios, we have collected data from GPS and cell tower using GPSLogger for android. The app can stream data in real-time. The indoor navigation data was gathered by a Ubisense, Inc. Ultra-wideband (UWB) radio research pack. Before the data is stored in database initial processing such as noise removal and compression are done. Initial processing on the data has been discussed in the paper ADDIN CSL_CITATION { "citationItems" : [ { "id" : "ITEM-1", "itemData" : { "ISBN" : "9781457702204", "author" : [ { "dropping-particle" : "", "family" : "Kumar", "given" : "Ashish", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Lau", "given" : "Chiew Tong", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Chan", "given" : "Syin", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Ma", "given" : "Maode", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Kearns", "given" : "William D", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" } ], "id" : "ITEM-1", "issued" : { "date-parts" : [ [ "2016" ] ] }, "page" : "5401-5404", "title" : "A Unified Grid-based Wandering Pattern Detection Algorithm", "type" : "article-journal" }, "uris" : [ "" ] } ], "mendeley" : { "formattedCitation" : "[7]", "plainTextFormattedCitation" : "[7]", "previouslyFormattedCitation" : "[7]" }, "properties" : { "noteIndex" : 0 }, "schema" : "" }[7].Server (Data storage and Analytics engine):We can divide the server into three major sub-components: Data Store: Data store is a central repository for the profile of the patient. It contains the user’s demographic information and various dementias/wandering-related test results. These records are updated based on suggestions by doctors. A stream of locomotion data along with contextual information is stored in corresponding patients’ profile. Locomotion data from the sensor can either be streamed in real time or in batches. To study the feasibility of the system, we have used batch processing. This method has been preferred as it is very easy to implement and preserves the computing power of the server. Sensors do not have to be connected to the server at all the time. API Engine: An Application Programming Interface (API) allows the client such as smartphone or desktop to make use of the functionality of user authentication and authorization, database query and call to important services. It provides a consistent, programmatic method for accessing a resource; it is a structured way of exposing algorithmic functionality to be used by client. We have identified key services which is important for the wandering management. We abstract the services from the internal implementation of the algorithm. The API calls are made using HTTP methods such as GET, POST, UPDATE, DELETE etc. If at any time there is a requirement to add new services, it can be easily done on the fly without affecting the older services. The analytics engine hosts the services for identification of wandering pattern, calculation of statistical features and analysis of trends. All services are handled through the API call from client to the server.Webhooks for Alert generation: One of the important aspects of wandering management is to be aware of the abnormal and critical situation of the patient. Services running on the server can alert of any discrepancy in situation or in cases of abnormality in the wandering pattern. It can generate alerts to caretaker and doctor and it may also suggest some clinical test such as MMSE to better understand the situation. The alerts can be implemented in the form of webhooks, these are URL address which a client can subscribe to and server can push the alerts and notification on this address.Clients:A host of devices such as desktop or smartphone can act as a client for generation of charts using API calls to the server or receiving prompts through webhooks. The client is authenticated, and access permission is granted based on the user privilege. This prevents the data abuse from unauthorized user. OAuth 2.0 can be used for users’ credential management. For the new client to connect to the server, profile for the client must be created by user having administrative privileged. Implementation and Analysis of resultsWe have conducted a preliminary experiment to ascertain the feasibility of the system. In this experiment, we aim to show some of the important aspects of wandering which can be generated from the analysis of the navigational data of the patient. These results can be generated on the client side through API call to the server. For this experiment, we have used the navigational data of single patients residing in day care facility over a period of one year. During daytime when in motion, tags transmitted x, y, and z coordinates in centimeters at 0.43 second intervals with a time-stamp. These data after initial processing is stored in the database. We identify the different pattern of navigation using the algorithm as described in the paper [7]. REF _Ref502662371 \h \* MERGEFORMAT Table 3.3 shows the breakdown of different patterns during the day. REF _Ref502662461 \h \* MERGEFORMAT Figure 3.12 shows the corresponding trend of these patterns. During the progress of day percentage of the direct pattern decreases whereas random patterns increases. Lapping and pacing are very small fractions of all the navigation motion and does not change significantly. In REF _Ref466922792 \h \* MERGEFORMAT Figure 3.13, we can see that trends in the navigational pattern have been plotted for one year. From trend-analysis, it has been found that there is no significant increase or decrease of any pattern over a period of one year. From this graph, it may be interpreted that there is no decline in the cognitive state of the person over one year.Table 3.3: Percentage of navigational pattern during time of day (ToD)Pattern\ToD00:08-11:5912:00-15:5916:00-19:5920:00-23:59Direct %72.1655.1451.1645.42Random %22.5139.9643.9549.05Lapping %05.0702.6002.9302.60Pacing %00.2502.3001.9602.93Figure 3.12: Trend in navigation patternDiscussionsWith our novel approach to the wandering pattern detection and development of a new set of the grid-based algorithm, it is now possible to detect wandering and to calculate the statistical features automatically in real-time. Contextual information about wandering such as place, time, and environmental factors which are likely to trigger wandering can be easily identified by intensity- analysis algorithm on these features.It has been found ADDIN CSL_CITATION { "citationItems" : [ { "id" : "ITEM-1", "itemData" : { "DOI" : "10.4017/gt.2011.10.2.004.00", "ISBN" : "1569-1101", "ISSN" : "15691101", "abstract" : "Objective - We hypothesized path tortuosity (an index of casual locomotor variability) measured by a movement telesurveillance system would be suitable for assisted living facility residents clinically diagnosed with dementia. Background We examined the relationship of dementia to path tortuosity and to movement speed and path length variability, both of which increase in dementia. Methods Daytime movements of 25 elders (19 female; 14 with dementia; average age 80.6) were monitored for 30 days using radio transponders measuring location with a maximum accuracy of 20 cm. After 30 days, the Mini Mental State Exam (MMSE) and Revised Algase Wandering Scale-Community Version (RAWSCV) were administered. Results Fractal Dimension (Fractal D), a measure of path tortuosity, correctly classified all but 2 residents with dementia; sensitivity 0.857, specificity 0.818 while the MMSE had 6 misclassifications, a sensitivity of 0.857 and a specificity of 0.727. Individual logistic regressions of dementia diagnosis on predictors MMSE and Fractal D were significant, but a logistic regression using both predictors found Fractal D marginally predictive of dementia (p=0.055) while the MMSE was not (p=0.168). Although significantly correlated with Fractal D, rate of travel and mean path distance were not predictive of dementia. Fractal D correlated negatively with overall MMSE (r= -0.44, n=25, p<0.05) but the relationship was mediated by MMSE Geographical Orientation items. Fractal D was unrelated to the RAWS-CV. Conclusions Telesurveillance-measured path tortuosity is greater in persons diagnosed with dementia. Persons with dementia have relatively more impaired spatial memory which is required for successful navigation. Application Automatic monitoring of direction, length and speed of unconstrained movements.", "author" : [ { "dropping-particle" : "", "family" : "Kearns", "given" : "William D.", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Fozard", "given" : "James L.", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Nams", "given" : "Vilis O.", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Craighead", "given" : "Jeffrey D.", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" } ], "container-title" : "Gerontechnology", "id" : "ITEM-1", "issue" : "2", "issued" : { "date-parts" : [ [ "2011" ] ] }, "page" : "90-102", "title" : "Wireless telesurveillance system for detecting dementia", "type" : "article-journal", "volume" : "10" }, "uris" : [ "" ] } ], "mendeley" : { "formattedCitation" : "[6]", "plainTextFormattedCitation" : "[6]", "previouslyFormattedCitation" : "[6]" }, "properties" : { "noteIndex" : 0 }, "schema" : "" }[6] that parameters such as Fractal-Dimension correlate significantly with the cognitive state of dementia patients. Trend-analysis on these parameters over a longer duration of time can be used to ascertain the efficacy of medication or procedure to restrict wandering and it can also help in deciding the proper intervention process. It will help in better decision-making capability using cause-effect analysis on each individual subject.Figure 3.13: Trend diagram over a yearConclusion and Future WorkThere has been no tool which can analyse the various factors which may lead to wandering. Analysis of these factors tailored to the individual cases can give more insight into the very nature of wandering. The current study procedures involve videotaping navigation episodes of the patients and later manually recording these episodes as wandering and non-wandering episodes by trained professionals. This procedure is very labor intensive, cumbersome, time-consuming, and prone to errors. We have shown in our previous work ADDIN CSL_CITATION { "citationItems" : [ { "id" : "ITEM-1", "itemData" : { "ISBN" : "9781457702204", "author" : [ { "dropping-particle" : "", "family" : "Kumar", "given" : "Ashish", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Lau", "given" : "Chiew Tong", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Chan", "given" : "Syin", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Ma", "given" : "Maode", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" }, { "dropping-particle" : "", "family" : "Kearns", "given" : "William D", "non-dropping-particle" : "", "parse-names" : false, "suffix" : "" } ], "id" : "ITEM-1", "issued" : { "date-parts" : [ [ "2016" ] ] }, "page" : "5401-5404", "title" : "A Unified Grid-based Wandering Pattern Detection Algorithm", "type" : "article-journal" }, "uris" : [ "" ] } ], "mendeley" : { "formattedCitation" : "[7]", "plainTextFormattedCitation" : "[7]", "previouslyFormattedCitation" : "[7]" }, "properties" : { "noteIndex" : 0 }, "schema" : "" }[7] that the entire process can be automated reliably with very high accuracy. We have also found that our algorithm can also detect the change in these patterns over time and prompt us when abnormal behavior is detected. In the next chapter, we test our system in real-time with more number of cases and examine the overall performance of the system across multiple individuals with different patient characteristics. This will also help in advancing the science of wandering management by giving researchers a tool to understand its nature and giving them flexibility in the design of their research processes. Hospitals, nursing home and assisted living facility would be the main beneficiaries of this research. Our primary aim is to provide a tool which can be reliably used by medical and data scientist to better understand the nature of wandering. ADDIN EN.SECTION.REFLIST A Data Analytic Study of WanderingThe aim of this chapter is to test the statistical tool discussed in Chapter 3, for analyzing indoor or outdoor wandering patterns and investigate the relationship with the cognitive measures such as Mini Mental State Examination (MMSE) ADDIN EN.CITE <EndNote><Cite><Author>Folstein</Author><Year>1975</Year><RecNum>176</RecNum><DisplayText>[92]</DisplayText><record><rec-number>176</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">176</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Folstein, Marshal F</author><author>Folstein, Susan E</author><author>McHugh, Paul R</author></authors></contributors><titles><title>“Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician</title><secondary-title>Journal of psychiatric research</secondary-title></titles><periodical><full-title>Journal of psychiatric research</full-title></periodical><pages>189-198</pages><volume>12</volume><number>3</number><dates><year>1975</year></dates><isbn>0022-3956</isbn><urls></urls></record></Cite></EndNote>[92] and the Revised Algase Wandering Scale – Community Version (RAWS-CV) ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2004</Year><RecNum>177</RecNum><DisplayText>[93]</DisplayText><record><rec-number>177</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">177</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author><author>Son, G-R</author><author>Beattie, Elizabeth</author><author>Song, J-A</author><author>Leitsch, Sara</author><author>Yao, Lan</author></authors></contributors><titles><title>The interrelatedness of wandering and wayfinding in a community sample of persons with dementia</title><secondary-title>Dementia and geriatric cognitive disorders</secondary-title></titles><periodical><full-title>Dementia and geriatric cognitive disorders</full-title></periodical><pages>231-239</pages><volume>17</volume><number>3</number><dates><year>2004</year></dates><isbn>1421-9824</isbn><urls></urls></record></Cite></EndNote>[93]. IntroductionIn Section REF _Ref503291413 \r \h \* MERGEFORMAT 4.2 the author describes the methods used for the two studies done on the navigational data. The first study (Section 4.3) done on the indoor and outdoor navigational data, test the reliability of our proposed methods based on parameters such as accuracy, sensitivity, specificity, and F-measures. In the second study (Section REF _Ref503291493 \r \h \* MERGEFORMAT 4.4) we study the statistical features in these patterns and its correlation with the measures of cognition. MethodsThe creation of an automated tool to extract direct, random, lapping and pacing movement measures from real-time location data simultaneously from multiple individuals permit a deeper understanding of wandering behavior because of increased sample size and enables more precisely targeted interventions, since subtle individual differences become more apparent in longitudinal datasets gathered over significant intervals of weeks or months. Presently there is no specialized tool capable of extracting wandering patterns from real-time tracking data.SubjectsThe study was performed with the help of 25 volunteers (19 females) at two Assisted Living Facilities (ALFs) in Tampa, Florida ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2011</Year><RecNum>8837</RecNum><DisplayText>[90]</DisplayText><record><rec-number>8837</rec-number><foreign-keys><key app="EN" db-id="avf5v0ptmz900ne2xem5999f2tzfas2pdzsd" timestamp="1487280418">8837</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, W.</author><author>Fozard, J. L.</author><author>Nams, V. O.</author><author>Craighead, J. D.</author></authors></contributors><titles><title>Wireless telesurveillance system for detecting dementia</title><secondary-title>Gerontechnology</secondary-title><short-title>Wireless telesurveillance system for detecting dementia.</short-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>90-102</pages><volume>10</volume><number>2</number><section>90</section><dates><year>2011</year><pub-dates><date>2011</date></pub-dates></dates><urls></urls></record></Cite></EndNote>[90]. The participants’ mean age was 81 (SD = 9.5) years and the mean MMSE score was 17.7 (SD = 8.0). Fourteen subjects were diagnosed with dementia, and all were capable of independent movement with or without assistive devices. Independent sample t-tests indicated no significant differences in mean age or MMSE scores of the participants at the two sites. All participants wore an Ubisense compact tag transponder on their wrist, which transmitted x, y, z-coordinate location information in meters with respect to a fixed origin in one corner of the room, coincident with a timestamp at 0.43 second intervals when in motion. The tag was worn only during daytime for 30 days that transmitted data only when in motion. The data is collected while performing an activity of daily living (ADL) in the shared living area of assisted living facility (ALF). Common activities performed are: Going to the dining areaHolding Conversation in a common areaWatching TelevisionGoing out of the facilityThe MMSE scores and Fractal D values for navigation are presented in REF _Ref503352744 \h \* MERGEFORMAT Table 4.3.ApparatusUbisense, Inc. Ultra-wideband real-time location system ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2014</Year><RecNum>166</RecNum><DisplayText>[94]</DisplayText><record><rec-number>166</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">166</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Webster, Paul</author><author>Jasiewicz, Jan M</author></authors></contributors><titles><title>Location aware smart watch to support aging in place</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><volume>13</volume><number>2</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>[94] ( REF _Ref504662293 \h Figure 4.1) is a compact tag wrist-worn transponder measuring 38 mm × 39 mm × 16.5 mm with a weight of 25 g. Using ultra-wideband (UWB) it can localize target to 15cm in 3D in real time. It is quite robust and can be used in variety of environment such as homes, hospitals, or industries. Full compact tag transponder specifications are presented in REF _Ref504662104 \h Figure 4.2. It was used to generate the archival positioning data. Four sensors were installed, one at each corner of a common living room that contained items for social and recreational activities and furniture. Figure 4.1: A Ubisense Compact TagA Belkin, Inc. Power of Ethernet 100 BaseT switch and seven shielded category 5e network cables comprised the sensor network, which transferred data to a Dell Inspiron model 1501 notebook computer. Ubisense 2.0 software on the computer was used to process and store sensor data [6].Figure 4.2: Full Specification of Ubisense Series 7000 Compact Tag Description of the monitored area Four Ubisense 2.0 sensors were installed at each corner of common use spaces at each ALF; (25.6m by 9.3m) and (23.8m by 9.4m). The common space interconnected dormitory wings, an exterior exit, and a dining room where all subjects ate their meals. The monitored space served two major functions: it was the conduit between the dormitory wings and dining room or exterior door and it served as a gathering place for recreational activities with sofas, tables, comfortable chairs, and a television set. The room was often rearranged to accommodate a daily event, such as a musical activity. The monitored space, contiguous with the room layout, was approximately rectangular. The locations of the sensors and the entrances to the monitored space are shown in REF _Ref503354299 \h \* MERGEFORMAT Figure 4.3. Figure 4.3: Floor plans for research site 1 (top) and 2 (bottom); sensor locations are at the vertices of the shaded regions and the origin is in the lower left; major divisions are 10m increments; individual participants appear as numbered ovalsProcedure for data collectionWe collected data from both indoor as well as outdoor environment. The procedures for the data collection is described below:Indoor navigation data: Tags were attached to the patient after the morning meal and medications and surrendered before retiring to their room. Daytime movement of 25 elders was monitored for 30 days using ultra-wideband sensor network using wireless transponder measuring location with an accuracy of 20 cm. Subjects were to wear the small tag during their walking hours for 30 days. The tag was very safe and would emit very weak radio waves. The data contain the Tag ID, timestamp and x, y, z co-ordinates. When in motion, the tags transmitted coordinates in centimeters once every 0.43 seconds relative to an origin set in one corner of the room. Outdoor navigation data: Figure 4.4: Data collection method for outdoor navigation using GPS sensor from mobileThe outdoor data was collected using an HTC One M7 internal GPS receiver set to maximum accuracy and 2Hz frequency using an opensource tool GPSLogger. The accuracy of the recording depended on the Phone GPS chip and the outdoor environment. We used an open area for our experiment and were able to maximize accuracy up to 2m. We observed some noise in GPS data due to the reflection of signals from buildings and cloud, however we were able to identify these spikes based on the relative speed variability between two consecutive paths. Smoothing was done to remove these noises using a moving point average method. This was also verified by plotting these paths. GPS data were obtained as latitude, longitude which was transformed into a reference coordinate system. This was important so that the algorithm can be applied to any data source irrespective of its origin. One of the main difference in data collected from UWB sensor (indoor) and GPS sensor (outdoor) was in terms of accuracy. This was mitigated by the selection of appropriate grid-size for the algorithm. The selection of grid size larger than the maximum accuracy of the device made sense so that noise from the data is naturally removed and navigation trajectory is not affected by noise. For our case, the GPS trace from the mobile device was recorded from one subject who was instructed to simulate various walking patterns in a pre-defined order and then in a random order. The subject was instructed to delay by 60 seconds consecutive patterns and later to execute the patterns in random order without a delay. Approximately 12,000 readings were recorded over 5 hours. The subject in this experiment had no symptoms of dementia. A sample trace of path is shown in REF _Ref503348220 \h \* MERGEFORMAT Figure 4.4.Data Structure: The navigational data collected over the period is represented by the data structure as shown in REF _Ref417264141 \h \* MERGEFORMAT Figure 4.5. Here, the navigational data is split into episodes which represent the ambulation part of the navigation. Most of the data analysis is done on this part of navigation and non-ambulation part is used to identify the point of interest, where subject stops to perform some activities. These points are important to discover the frequent pattern in navigation. Dataset|_____Patient-1|_____Patient-2___________|||||_____Patient-k||||||| ……… |Date-1Date-2 Date-3 Date-m|||_______Episode-1|_______Episode-2||_______ Episode-i||_______Episode-nFigure 4.5: Data structure used for episode classificationWandering Pattern IdentificationIn this part of our study, we are categorizing the wandering pattern into four categories as defined by the Martino-Saltzman classification based on the geometrical pattern. This follows the workflow we described in the Chapter 3. Here, first we discuss the procedure on the raw data and finally, we evaluate the results in terms of algorithm accuracy.ProcedureThe pre-processing step was used to remove outliers and inconsistent data from the readings; exceptionally high-speed steps were identified as outliers and removed as noisy data. The noise may have been caused by reflected signals from buildings or metal structures in the ALF (for UWB). Due to the high sampling rates of the devices as well as subject inactivity (e.g. halting or rarely moving), data collected, especially from UWB, were found to be very crowded and not all points were required to plot an episode. The algorithm processing speed was adversely affected by these large number of points. We used a mean shift clustering algorithm to reduce the number of points. Episode segmentation based on the stopping criteria of 60 seconds performed very poorly in the real-life situation since even when there was no physical movement, sensors transmitted new trace information due to noise. The author employed smoothing followed by filtration based on speed to segment navigation from non-navigation. This resulted in very good segmentation accuracy in both indoor and outdoor scenarios.Verification of algorithm using controlled experimentTo verify the algorithm performance, a controlled experiment was conducted using outdoor navigation data as explained in the Section: REF _Ref522563735 \r \h 4.2.4.2. The subject was given instruction to trace the patterns in any sequence and span the distance of more than 10 meters for each pattern. The subject was also instructed to halt for 60 seconds before tracing the new pattern. 100 patterns were recorded, 25 each direct, random, lapping and pacing. Data collected for this experiment has a well-defined start and end timestamp along with the coordinate information for each pattern along with the label of the pattern. The grid-size parameter was selected to be 2 meters, so that path shape is not affected by the noise in the data. Classification result is shown in REF _Ref522563799 \h Table 4.1. The algorithm was able to identify direct and random pattern quite reliably with the accuracy of 92% each. While some of the lapping and pacing episodes were misclassified. The reason for lapping to be misclassified as pacing and vice-versa were mainly due to the fact that distinction was based on the area enclosed by the loop which is a rather subjective and depends on the human judgement for the distinction between these patterns.Table 4.1: Classification Table for the controlled experiment?Predicted?DirectRandomLappingPacingTotalsActualDirect2320025Random0231125Lapping0121325Pacing0222125Totals23282425100ResultsA total of 823 episodes were extracted from the indoor navigation data, but most of these episodes were either direct or random and very few were identified as lapping or pacing. To tackle the class imbalance problem, we selected equally sized classes of 25 episodes per class from the UWB data. This was matched with equal number of cases per class from the outdoor navigation data. 200 distinct episodes for testing of our proposed algorithm was used, 100 each from indoor and outdoor navigation. To ascertain the accuracy of identification, ground truth for the pattern was established through manual plotting and labelling of these episodes. All the episodes were processed in the same manner irrespective of originating in an indoor or outdoor environment. REF _Ref502437220 \h \* MERGEFORMAT Figure 4.6 shows the recall and precision of our proposed algorithm. For the direct pattern, recall (94%) and precision (98%) was found to be very good and the algorithm was able to identify almost all instances with very few misses. REF _Ref504667543 \h \* MERGEFORMAT Table 4.2 shows the classification table. Actual episodes are shown in rows and column contains the predicted output. The random pattern also has a very high recall value (92%), but the precision (85%) was lower because the other three patterns were occasionally misclassified as a random pattern. Lapping and pacing also performed well, but precision and recall were a bit lower, particularly because of the complex nature of these patterns. Some of lapping patterns were misclassified as pacing and vice-versa. Overall the algorithm was successful in classifying most of the patterns and had an accuracy of 90%.Table 4.2: Classification Table?Predicted?DirectRandomLappingPacingTotalsActualDirect4730050Random1461250Lapping0244450Pacing0344350Totals48544949200Figure 4.6: Precision and recall of the proposed algorithmDiscussionThere has been previous attempt to classify these patterns but it is difficult to do a fair comparison with other similar methods, the closest being an algorithm proposed by N. K. Vuong et al. ADDIN EN.CITE <EndNote><Cite><Author>Vuong</Author><Year>2014</Year><RecNum>130</RecNum><DisplayText>[78]</DisplayText><record><rec-number>130</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">130</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Vuong, NK</author><author>Chan, S</author><author>Lau, CT</author></authors></contributors><titles><title>Automated detection of wandering patterns in people with dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>127-147</pages><volume>12</volume><number>3</number><dates><year>2014</year></dates><isbn>1569-111X</isbn><urls></urls></record></Cite></EndNote>[78], where they compared sensitivity, specificity, precision, recall and F1-measure for different deterministic and machine learning algorithms. But their dataset used room-to-room movement information of subjects, whereas we used fine-grained coordinates of the physical layout, which is most generic form of data collected from sensors in indoor localization.On average, walking stride length is 2.5 feet or 75 cm. For indoor navigation, UWB sensor can capture every stride and hence the exact path trajectory, but in the case of outdoor navigation since the accuracy was much larger than average stride length it was difficult to capture exact path trajectory but for the kind of analysis we were doing there was a negligible loss of path information. For the cases where the patterns were smaller than 2m, it was not possible to identify it in case of outdoor navigation. After selection of grid-size the processing steps are same which does not differentiate between these two different sources of data.Statistical analysis of the Wandering Pattern In the previous section, we have shown that our algorithm has 90% classification accuracy and is applicable to indoor as well as outdoor navigation data. For the next part, we investigate the statistical features of the episodes and its correlation with cognitive measures such as MMSE and RAWS-CV. Procedure Matlab R2013a was used for all data processing tasks and implementation of the algorithm. Total distance travelled, and time of ambulation was calculated for each patient over a period of 30 days and the fractional time and distance accrued in each pattern (direct, random, lapping and pacing) was calculated as shown in REF _Ref503352744 \h \* MERGEFORMAT Table 4.3. Approximately 12,000 hours of navigational data were collected from the two ALF sites. This generated approximately 2.6M unique tag position sightings. The number of sightings was reduced to 699,205 (27%) after initial processing and partitioned into ambulatory and sedentary phases. Total distance travelled ranged from 2 km to 102 km with a mean of 22 km and median of 15.18 km. There were 6 subjects who travelled less than 10 km. Travel speed ranged from 0.10 m/s to 0.52 m/s with a mean of 0.24 m/s (SD=0.10) and a median of 0.21 m/s. IBM SPSS Ver. 20 was used to calculate descriptive and inferential statistics. There were very few cases of missing data and in those instances missing observations were replaced with the sample mean.Results The 12,905 ambulatory episodes were coded using the proposed method. Several episodes with travel durations under 120 seconds were discarded. Episodes very near or adjacent in temporal domain (120 seconds apart or less) were merged resulting in 4,030 episodes that were categorized into the four navigational typologies. Two parameters, the distance traversed, and the time spent in each navigational pattern were used to estimate the fractional distribution of navigational patterns across both groups ( REF _Ref503352744 \h \* MERGEFORMAT Table 4.3). The distribution of the patterns based on distance travelled within each was: Direct (mean = 32.57%, SD = 14.56), Random (mean = 59.89%, SD = 13.18), Lapping (mean = 4.02%, SD = 3.19), and Pacing (mean = 3.51%, SD = 2.43). The distribution of time spent within each pattern was: Direct (mean = 21.85%, SD = 13.30), Random (mean = 70.79%, SD = 12.09), Lapping (mean = 3.26%, SD = 4.69), and Pacing (mean = 4.09%, SD = 3.00). Since both the temporal and distance distributions were very similar, we present the fractional distribution of the distance travelled within each pattern (a spatial analysis) for the remainder of the analyses unless otherwise stated. As noted, random and direct were the most frequent travel patterns, whereas lapping and pacing constituted a very small percentage; results which concur with an earlier study by Algase ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>1997</Year><RecNum>168</RecNum><DisplayText>[14]</DisplayText><record><rec-number>168</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">168</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, Donna L</author><author>Kupferschmid, Barbara</author><author>Beel-Bates, Cynthia A</author><author>Beattie, Elizabeth Ra</author></authors></contributors><titles><title>Estimates of stability of daily wandering behavior among cognitively impaired long-term care residents</title><secondary-title>Nursing Research</secondary-title></titles><periodical><full-title>Nursing research</full-title></periodical><pages>172-178</pages><volume>46</volume><number>3</number><dates><year>1997</year></dates><isbn>0029-6562</isbn><urls></urls></record></Cite></EndNote>[14] in which lapping and pacing were very infrequent events. Independent sample t-tests found no differences in the breakdown of patterns between the two ALF sites. The presence of random, lapping and pacing characterizes wandering in dementia. By extension those without a diagnosis of dementia should present a reduced proportion of wandering patterns and a higher proportion of direct travel. An independent groups t-test of the algebraic sum of the distances traversed in the wandering patterns (random, lapping and pacing) by subjects with diagnoses of dementia in their ALF’s clinical record vs. controls was significant (t = -2.40, df = 23, p < 0.05). The percentage of distance travelled in all wandering patterns in the dementia group was 73.08% (SD = 8.46) vs. 60.23% (SD = 17.73) for subjects with no diagnosis of dementia in their records.Table 4.3: Fractional distribution of navigational pattern in subjects staying in ALFRelationship of MMSE Spatial and Temporal Orientation to the wandering pattern: Within the MMSE, two items, i6f and i7f measure temporal orientation and spatial orientation respectively, with low scores indicating greater disorientation. Using each items’ midpoint of 3 (range 1 to 5) to bisect the sample, 9 subjects showed good temporal orientation (mean = 4.33, SD = 0.83) and 16 subjects (mean = 1.00, SD = 0.89) had poor temporal orientation. Similarly, for spatial orientation, 13 subjects showed good orientation (mean = 4.38, SD = 0.87) and 12 subjects indicated poor spatial orientation (mean = 1.17, SD = 0.93). An independent sample t-test found percentage distance lapping significantly discriminated among the MMSE temporal orientation groups (i6f) (t = -2.48, df = 23, p < 0.05; mean = 2.95%, SD = 2.81 vs. 5.94% SD=3.06) with better temporal orientation associated with more lapping. Percent distance travelled in random and pacing did not differentiate the temporal orientation subgroups. In contrast, for spatial orientation (i7f) subgroups, spatially disoriented subjects showed only increased random travel compared to spatially oriented subjects (t = 3.50, df = 23, p < 0.01; mean = 67.81% SD=8.15 vs. 52.58% SD = 12.88 vs.).Logistic regression to identify dementia diagnostic group:A binary logistic regression was conducted to predict dementia diagnosis using only percentage distance traversed in each of the three wandering patterns (random, lapping, and pacing) as predictors. In the forward selection model, random (B=0.116, df=1 p=0.03) and lapping (B= 1.50, df=1 p=0.073) travel measures predicted dementia status, with an overall accuracy=80.0%, Sensitivity=78.6% and Specificity=81.8%). The Homer and Lemeshow test performed to check goodness-of-fit for this group of 25 subjects showed no lack of fit (Chi-square=4.694, df=6, Sig=0.584). By comparison, Kearns et al. ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2011</Year><RecNum>162</RecNum><DisplayText>[24]</DisplayText><record><rec-number>162</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">162</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Nams, Vilis O</author><author>Craighead, Jeffrey D</author></authors></contributors><titles><title>Wireless telesurveillance system for detecting dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><volume>10</volume><number>2</number><dates><year>2011</year></dates><urls></urls></record></Cite></EndNote>[24] employed Fractal-D as predictor of dementia status and achieved slightly better results (B = 26.61, df = 1, p = 0.007; Accuracy=84%, Sensitivity=82% and Specificity=86%). Table 4.4: Classification TableContrast of the proposed algorithm vs. Fractal Dimension:Kearns et al. contend that the amount of random variation in a subject’s movement path, measured as Fractal Dimension, is associated with level of spatial confusion, and have published several substantiating articles. To date Fractal Dimension has not been contrasted with Martino-Saltzman’s typology. To determine the association of these approaches we computed Pearson product moment correlation coefficients between the percentage distances traversed in each of Martino-Saltzman’s categories vs. the Fractal Dimension value for each subject averaged from their movement paths. We have calculated the correlation table with respect to spatial ( REF _Ref504667925 \h Table 4.5) as well as temporal ( REF _Ref504667942 \h Table 4.6) parameter. The results displayed in REF _Ref504667925 \h Table 4.5 and REF _Ref504667942 \h Table 4.6 reveal Fractal Dimension is strongly and inversely related to direct travel, and positively correlated with random and pacing travel patterns, but not significantly associated with lapping behavior. Here, N is total no of subjects and Sig is the significance of test variable. The test has been performed at 0.01 significance level. REF _Ref504667925 \h Table 4.5 and REF _Ref504667942 \h Table 4.6 compare the temporal measures and spatial measures against Fractal D, the MMSE and the wandering status of the subjects. The results give a great insight into Fractal D relation to the Martino-Saltzman’s typology. ? Table 4.5: Correlations of Fractal D and MMSE with Spatial Measures of Wandering?Fractal DMMSEWandererDirectRandomLappingPacingFractal D1-.442*0.372-.846**.859**-.459*.724**MMSE?1-0.3650.360-.432*0.305-0.076Wanderer??1-0.3170.340-0.2770.290Direct???1-.974**0.279-.773**Random????1-.458*.690**Lapping?????1-0.276Pacing??????1*. Correlation is significant at the 0.05 level (2-tailed).**. Correlation is significant at the 0.01 level (2-tailed).Table 4.6: Correlations of Fractal D and MMSE with Temporal Measures of Wandering?Fractal DMMSEWandererDirectRandomLappingPacingFractal D1-.442*0.372-.821**.821**-.535**.613**MMSE?1-0.365.440*-.508**0.368-0.087Wanderer??1-0.1910.206-0.3010.230Direct???1-.977**.480*-.699**Random????1-.596**.591**Lapping?????1-.430*Pacing??????1*. Correlation is significant at the 0.05 level (2-tailed).**. Correlation is significant at the 0.01 level (2-tailed).Discussion:A paradigm shift in dementia research and care is providing clinicians and researchers with new instruments that will lead to a fuller comprehension of its origin and treatment. Technologies directed at wandering have to date been largely preventive ADDIN EN.CITE <EndNote><Cite><Year>2017</Year><RecNum>196</RecNum><DisplayText>[95, 96]</DisplayText><record><rec-number>196</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">196</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors></contributors><titles><title>Vivago develops smart safety and wellbeing solutions for preventive care</title><short-title>Vivago develops smart safety and wellbeing solutions for preventive care</short-title></titles><keywords><keyword>elderly care, safety, wellbeing, preventive care</keyword></keywords><dates><year>2017</year></dates><urls><related-urls><url> app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">197</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors></contributors><titles><title>Offender Monitoring | Numerex IoT Solutions</title><short-title>Offender Monitoring | Numerex IoT Solutions</short-title></titles><dates><year>2017</year></dates><urls><related-urls><url>;[95, 96]; restricting wandering by generating alerts or initiating behavioral interventions when wandering was detected, or tracking down an individual who may have eloped. Recent improvements in analytic tools have shifted the momentum from restricting to managing wandering, perhaps due to an emerging contention that wandering may provide opportunities for relieving stress, discharging excess energy, and obtaining healthful physical exercise. Our approach for more precisely categorizing wandering is aligned with this goal. A plethora of inexpensive devices are being marketed capable of generating significant amounts of movement tracking data at various levels of resolution. An increasing number of these technologies can generate frequent sightings in three dimensions such as GPS (outdoor) and Ultra-wideband (indoor). Despite the presence of these 3D tracking technologies, research on wandering has suffered from a dearth of statistical tools tailored to extracting meaningful results from raw tracking information. The result has been that the gathered information seldom realizes its full potential. In this chapter, we presented a tool to analyse wandering behavior and categorize it automatically into Martino-Saltzman’s wandering pattern schema of direct travel, random, lapping or pacing. We have also contrasted it with Fractal Dimension; a spatial measure of randomness in a person’s movement-path that is moderately correlated with Mini Mental State Exam scores, dementia diagnosis, and fall likelihood and determined that the Fractal Dimension is strongly and positively associated with random and pacing travel but not lapping. Lapping, although a relatively infrequent event was paradoxically more likely to be present in subjects with no diagnosis of dementia. Prior research that has dichotomized movement data into direct and random patterns has resulted in little advancement in the understanding of relatively infrequent lapping and pacing behaviour. Increased availability of this new tool should yield a deeper understanding of these wandering subtypes and their relationship to clinical measures such as the MMSE.ConclusionTo our knowledge, this study represents the first attempt to automatically classify all four navigation patterns using Martino-Saltzman classification on a longitudinal dataset containing person-level tracking data. As predicted, a random travel pattern was significantly more prevalent in subjects having low MMSE scores and diagnoses of dementia. Surprisingly, lapping was more frequent in the group with less cognitive impairment. No significant relationships were found between pacing episodes and any of the clinical measures.We examined two subtypes of wandering: lapping and pacing, for their relationship to MMSE measures of temporal and spatial orientation. Lapping was strongly correlated with the temporal orientation but not spatial orientation. Pacing was unrelated to either spatial or temporal orientation. Analysis of trend in the navigationIn the last two chapters, the author has demonstrated the design and implementation of the algorithm to identify the wandering patterns on the real-world data set of 25 subjects living in an assisted living facility. The dataset for each subject was for a relatively short duration of 30 days. It was sufficient for the type of analysis the author performed in the previous chapters. However, for a longitudinal study on the navigational features, one-month data is not sufficient to reliably uncover any trend. To overcome this short-comings the author has investigated the data over a longer period. In this chapter the author analyse the data for a better understanding of the progression of dementia in ALF residents.IntroductionIn this chapter, the author analyses the data of 10 subjects collected from the ALF facilities in USA for approximately one year and examine the features of navigation to discover any trend or correlation with the test of dementia. We assume that if there is any consistent change in the wandering condition it should be observable over this long duration. In a similar study of the descriptive parameter, path-tortuosity has been identified as the significant indicator of cognitive impairment ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2010</Year><RecNum>180</RecNum><DisplayText>[15]</DisplayText><record><rec-number>180</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">180</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Nams, VO</author><author>Fozard, James L</author></authors></contributors><titles><title>Tortuosity in movement paths is related to cognitive impairment</title><secondary-title>Methods Inf Med</secondary-title></titles><periodical><full-title>Methods Inf Med</full-title></periodical><pages>592-598</pages><volume>49</volume><number>6</number><dates><year>2010</year></dates><isbn>0026-1270</isbn><urls></urls></record></Cite></EndNote>[15]. Path-tortuosity is defined in terms of Fractal Dimension (Fractal D), which varies from a value of 1, representing a perfectly straight path, to a value of 2 indicating a completely random path. Fractal D captures spatial variability in terms of the geometry of the path and is employed to study changes in path direction. Fractal D has augmented earlier research findings by elucidating the severity and trajectory of cognitive deficit ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2016</Year><RecNum>134</RecNum><DisplayText>[79]</DisplayText><record><rec-number>134</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">134</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Nams, Vilis O</author></authors></contributors><titles><title>Movement Path Tortuosity in Free Ambulation: Relationships to Age and Brain Disease</title><secondary-title>IEEE Journal of Biomedical and Health Informatics</secondary-title></titles><periodical><full-title>IEEE Journal of Biomedical and Health Informatics</full-title></periodical><dates><year>2016</year></dates><urls></urls></record></Cite></EndNote>[79]. MotivationDuring dementia’s progression, the patient typically undergoes a series of clinical evaluations. A physician who prescribes medication to address behavioral symptoms such as agitation and confusion is keenly interested in the effectiveness of that medication on linked aberrant behaviors such as wandering. The perceived effectiveness of the medication may determine the future course of action, indicating a dosage change or a substitute therapy. Tracking small but observable changes in behavior related to medication over long durations pose technical challenges but may be preferable for enacting timely interventions; quantifiable behavioral measures are preferable to less reliable first-person observations and hunches. The development of such a system would allow clinicians and researchers to create day-to-day charts of wandering parameters. The system could be used to identify abnormal deviations in these parameters, and indicate a test for cognitive impairment such as Mini Mental State Examination (MMSE) ADDIN EN.CITE <EndNote><Cite><Author>Folstein</Author><Year>1998</Year><RecNum>7308</RecNum><DisplayText>[97]</DisplayText><record><rec-number>7308</rec-number><foreign-keys><key app="EN" db-id="avf5v0ptmz900ne2xem5999f2tzfas2pdzsd" timestamp="1487280407">7308</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Folstein, S.</author></authors></contributors><titles><title>Key Papers in Geriatric Psychiatry. Mini-Mental State: A Practical Method for Grading the Cognitive State of Patients for the Clinician. M. Folstein, S. Folstein and P. McHugh, Journal of Psychiatric Research (1975) 12, 189_198</title><secondary-title>International Journal of Geriatric Psychiatry</secondary-title><short-title>Key Papers in Geriatric Psychiatry. Mini-Mental State: A Practical Method for Grading the Cognitive State of Patients for the Clinician. M. Folstein, S. Folstein and P. McHugh, Journal of Psychiatric Research (1975) 12, 189_198.</short-title></titles><periodical><full-title>International Journal of Geriatric Psychiatry</full-title></periodical><pages>285-294</pages><volume>13</volume><dates><year>1998</year><pub-dates><date>1998</date></pub-dates></dates><label>478</label><urls></urls></record></Cite></EndNote>[97] or the Revised Algase Wandering Scale - Community Version (RAWS-CV) ADDIN EN.CITE <EndNote><Cite><Author>Algase</Author><Year>2004</Year><RecNum>7879</RecNum><DisplayText>[98]</DisplayText><record><rec-number>7879</rec-number><foreign-keys><key app="EN" db-id="avf5v0ptmz900ne2xem5999f2tzfas2pdzsd" timestamp="1487280410">7879</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Algase, D. L.</author><author>Son, G. R.</author><author>Beattie, E.</author><author>Song, J. A.</author><author>Leitsch, S.</author><author>Yao, L.</author></authors></contributors><auth-address>University of Michigan School of Nursing, Ann Arbor, MI 48109-0482, USA. dalgase@umich.edu</auth-address><titles><title>The interrelatedness of wandering and wayfinding in a community sample of persons with dementia</title><secondary-title>Dementia & Geriatric Cognitive Disorders</secondary-title><short-title>The interrelatedness of wandering and wayfinding in a community sample of persons with dementia</short-title></titles><periodical><full-title>Dementia & Geriatric Cognitive Disorders</full-title></periodical><pages>231-9</pages><volume>17</volume><number>3</number><dates><year>2004</year><pub-dates><date>2004</date></pub-dates></dates><label>19</label><urls></urls></record></Cite></EndNote>[98] is required when such deviations occur thereby increasing the possibility of timely intervention by clinicians and caretakers.Visual inspection methods have limitations in calculating navigational parameters descriptive of wandering such as randomness, average speed, path-length etc. A computational method using the mathematical representation of the path is best suited to identify and quantify these features. We aim to calculate such features from navigation data and highlight their analytic value and identify their strength and importance in distinguishing two diagnostic groups. We will also examine linear trends in the features over a lengthy interval, which yields useful information on dementia’s progression.MethodsThe Proposed AlgorithmThe first two stages: data pre-processing and segmentation are identical to the steps we have discussed for wandering pattern detection in the previous chapter. In the third stage, we calculate important features and carry on the statistical analysis to discover trend if any. The features generated for this analysis has been described in detail in Section REF _Ref504673109 \r \h 5.3.2. REF _Ref502767721 \h \* MERGEFORMAT Figure 5.1, shows the complete processing steps, which consists of three stages. In the Data pre-processing stage, data collected from the UWB sensor array contains significant amounts of sensor noise. We cleaned the data before starting a mining task; the data is first passed through outlier detector using a Heuristic-based outlier detection method. One category of outliers removed is location data lying outside the physical layout of the facility. The other outlier category identified is based on the mean navigation speed for preceding points for a fixed window length. If the speed exceeds a threshold value it is considered a potential outlier. The next segmentation phase partitions the data into ambulation and sedentary (no subject movement) episodes. One challenge associated with identifying sedentary episodes concerns sensor noise, which can create false indications of movement. Empirical evaluation indicates noisy data vary significantly. From their neighbors in terms of travel speed and sharp turns made by successive points. These considerations have helped us reduce noise substantially. Sedentary episodes hold importance for identifying the Point of Interest (POI) in the physical layout. In the final phase of our processing scheme, features are calculated for the ambulation episodes, and trend-analysis using linear contrast analysis is performed to discern trends in monthly mean values of these parameters.Figure 5.1: The Complete FrameworkFeatures GenerationWe define the navigational pattern as the spatio-temporal sequence in some geographical spaces, represented by a series of chronologically ordered points called burst of location as Bi, for example, B1 → B2 → ... → Bn. Where each point represents spatial coordinates and time stamp of the point as B = (x, y, t), shown in REF _Ref502765985 \h \* MERGEFORMAT Figure 5.2. The categories of features that can be extracted from these sequential data are:SpatialTemporalSpatio-temporalContextualFigure 5.2: A sample trajectory pathExamples of spatial features are path-tortuosity, angle turn per unit distance, mean episode-length, etc. Temporal features include: time of navigation for each day, and the fractional distribution of time for each wandering pattern. Spatio-temporal features involve interactions of both domains, yielding heat-maps of the wandering pattern within the environment at different times of the day, and the intensity of different wandering patterns in the context of varying external or internal stimuli. Spatio-temporal data contain state information of the navigation, an event, or a position in space over a period of time. It poses many challenges to representing, processing, analyzing, and mining such datasets due to the complex structures of spatio-temporal objects, and the relationships among them. REF _Ref502765985 \h \* MERGEFORMAT Figure 5.2 represents a sample episode. The subject traverses the path and makes many turns before reaching its destination. The parameter that can be used to uniquely characterize the subject’s trajectory is Burst, it contains the spatio-temporal information about the path. All the trajectory features can be derived from this information. The sample path consists of six bursts B1 to B6; d1 to d5 is the distance in meters between each pair of bursts and α1 to α5 are the absolute angle in degrees [0,180] irrespective of clockwise or anti-clockwise direction.We define:The descriptive features are calculated for each subject using the formula as defined below:Along with aforementioned features we calculate the fractional duration of wandering pattern based on distance traversed and time spanned for each pattern as shown in REF _Ref503643633 \h \* MERGEFORMAT MMSEGroupNo of subjectsMeanSDClinical diagnosis of dementia613.337.6No dementia4199Table 5.2.Trend Analysis in NavigationThe questions:“The features such as walking speed, path efficiency, path tortuosity, etc. can be easily calculated from navigational data of the ALF residents. Do these features vary over lengthy period (nearly 12 months) for the subjects and, how does it differ among two groups – PwD and non-PwD?” “Is there any significant difference in trend between the two different groups of people - PwD and non-PwD?”Procedure: We use one-way ANOVA to compare means of the features, grouped in months in chronological order as shown in REF _Ref504686852 \h Table 5.1. We use linear contrast analysis to find significance of test. Additionally, we performed Tukey’s test and Games Howell test to ascertain homogeneity of variance assumption. The factor/features we have considered for analysis are: Fraction of ambulation episode (amb-frac): This is fraction of time person ambulates for each day. amb-frac = time of ambulation/total time of recordingRecording time: It roughly starts in the morning and ends by midnight. Speed (speed-amb): This is average speed of the ambulation episode for each dayPath efficiency (path-eff): Path efficiency of the person for each dayAngle turn per unit distance (angle-turn): Angle turn per unit distance for each day.For comparison of mean of the features we have following null hypothesis:Null Hypothesis: There is no change in the navigational features over the period.35627082846720377417686995Features00FeaturesTable 5.1: Grouping of features in chronological order2211245219653Group: 20Group: 21639022160881Group: 10Group: 1DATASET ACTIVATE DataSet#.UNIANOVA ang BY Mon /CONTRAST(Mon)=SPECIAL(-11 -9 -7 -5 -3 -1 1 3 5 7 9 11) /METHOD=SSTYPE(3) /INTERCEPT=INCLUDE /PRINT=ETASQ DESCRIPTIVE /CRITERIA=ALPHA(.05) /DESIGN=Mon.We have performed the trend analysis in IBM SPSS statistical toolbox 20.0, the sample code is: Here, significance value used is 0.05 to identify significant trend, denoted by parameter Alpha. We hypothesized that linear trends in the features over lengthy intervals might provide useful information concerning dementia’s progression. We employed linear contrast analysis to identify increasing and decreasing trends in the features and evaluated the reliability of the change using a one-way ANOVA to compare the trends for the two diagnostic groups.Subject and LayoutAn archived indoor position-tracking data set recorded from 10 volunteers in an Assisted Living Facilities (ALFs) in Tampa, Florida, provided the source of the archival information, which has been described in detail elsewhere ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2011</Year><RecNum>136</RecNum><DisplayText>[87]</DisplayText><record><rec-number>136</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">136</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Nams, Vilis O</author><author>Craighead, Jeffrey D</author></authors></contributors><titles><title>Wireless telesurveillance system for detecting dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>90-102. doi: 10.4017/gt. 2011.10. 2.004. 00</pages><volume>10</volume><number>2</number><dates><year>2011</year></dates><urls></urls></record></Cite></EndNote>[87]. Six subjects had received clinical diagnoses of dementia with MMSE scores averaging 13.33 (SD=7.6) while the four control subjects MMSE averaged 19 (SD=9), as shown REF _Ref522628205 \h Table 5.2. All were capable of independent movement with or without assistive devices. All participants wore a Ubisense ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2014</Year><RecNum>151</RecNum><DisplayText>[94]</DisplayText><record><rec-number>151</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">151</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Fozard, James L</author><author>Webster, Paul</author><author>Jasiewicz, Jan M</author></authors></contributors><titles><title>Location aware smart watch to support aging in place</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><volume>13</volume><number>2</number><dates><year>2014</year></dates><urls></urls></record></Cite></EndNote>[94] compact tag during daytime for approximately one year which transmitted data only when in motion. The tag was worn on the wrist, which transmitted x, y, z-coordinate location information in meters with respect to a fixed origin in the corner of the room, coincident with a timestamp at 0.43 second intervals. The monitored space was approximately rectangular with dimension: 25.6m X 9.3m. Approximately 7.7 million observations were generated during this period. REF _Ref504687047 \h Table 5.3 shows the subject demographics along with wandering pattern distribution.MMSEGroupNo of subjectsMeanSDClinical diagnosis of dementia613.337.6No dementia4199Table 5.2: Subject demographic Table 5.3: Distribution of wandering patternResults and DiscussionTrend AnalysisTwo subjects (Subject-1 and Subject-2) evinced significant linear trends in angle-turn and path-eff with the maximum variability captured by angle-turn (14.7 and 11.7). The trend is shown in REF _Ref502768039 \h \* MERGEFORMAT Figure 5.3. Both subjects were later found to have very low MMSE value (6 and 9 respectively). The plot in REF _Ref502768039 \h \* MERGEFORMAT Figure 5.3 depicts the trend observed in both subjects over one year. A decreasing trend was observed in path efficiency (path-eff) for both the subjects and increasing trend was observed for angle turn per unit distance (angle-turn) parameter. Trend observed for ambulation fraction (amb-fract) and average speed (speed-amb) was not significant. In four other residents angle-turn consistently increased over the 1-year monitoring interval suggesting that their cognitive abilities may have correspondingly deteriorated over this interval. Subject 9 also has very low MMSE score of 0, but there was no significant trend observed in it. It also has very high fraction or random pattern consistently over the time.Figure 5.3: Trend in features over a 1-year durationDistinction between two diagnostic groups: Two groups are: 6 subjects with a clinical diagnosis of dementia and, 4 control subjects. We calculated the features associated with each group. It can be seen from the box plot ( REF _Ref503651468 \h \* MERGEFORMAT Figure 5.4) that there is a significant difference between two groups as expected. Subjects who have dementia depicted low path efficiency, more angle turn/unit distance.Figure 5.4: Box plot for Dementia and Control GroupConclusionThe method presented can identify features that may reliably differentiate older adults with diagnoses of dementia and those without. We have devised the method for quantifying movement variability from real-time data acquisition methods and employed it to study movement related cognitive decline in ALF residents. For example, those whose dementia worsens over time will likely evince increasing angle-turn and decreasing path-eff over long durations. In this pilot investigation, we identified small changes in travel-speed, path-efficiency, angle-turn and, ambulation-fraction also associated with decreased cognitive function. We were able to reliably identify trends in the data that were later supported by a low MMSE score in the subject. In the next chapter we examine a larger group of subjects with a larger feature set to comprehensively analyse the strengths and weaknesses of this methodology. Nevertheless, this study has yielded new insights into dementia-related wandering.Summary and Future work ADDIN EN.SECTION.REFLIST While the number of studies related to wandering has increased in recent years, many gaps in science remain, limiting the empirical evidence on which to base important clinical decisions. These gaps in science contribute to significant variation in practice associated with assessment practices for wandering as well as interventions used to manage wandering. In this chapter, the author summarizes the research carried on as well as sets the direction for the future work.A change in basic assumptions in dementia research and care is providing clinicians and researchers with new instruments that will lead to a fuller comprehension of its origin and treatment. Technologies directed at wandering have to date been largely preventive; restricting wandering by generating alerts or initiating behavioral interventions when wandering was detected or tracking down an individual who may have eloped. Recent improvements in analytic tools have shifted the momentum from restricting to managing wandering, perhaps due to an emerging contention that wandering may provide opportunities for relieving stress, discharging excess energy, and obtaining healthful physical exercise. Our approach for more precisely categorizing wandering is aligned with this goal. Early research establishing wandering typologies was of necessity observational in nature; patients’ actions were manually coded and in later studies behaviors were videotaped. The dominant wandering pattern present was captured using predetermined coding sequences ADDIN EN.CITE <EndNote><Cite><Author>Martino-Saltzman</Author><Year>1991</Year><RecNum>159</RecNum><DisplayText>[7]</DisplayText><record><rec-number>159</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1514135579">159</key><key app="ENWeb" db-id="">0</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Martino-Saltzman, David</author><author>Blasch, Bruce B</author><author>Morris, Robin D</author><author>McNeal, Lisa Wynn</author></authors></contributors><titles><title>Travel behavior of nursing home residents perceived as wanderers and nonwanderers</title><secondary-title>The Gerontologist</secondary-title></titles><periodical><full-title>The Gerontologist</full-title></periodical><pages>666-672</pages><volume>31</volume><number>5</number><dates><year>1991</year></dates><isbn>0016-9013</isbn><urls></urls></record></Cite></EndNote>[7]. This required considerable time and effort by highly trained observers and constituted an extended vigilance task prone to human error so typically only a very few subjects could be observed simultaneously. Algase found her observers were capable of reliably detecting instances of lapping and pacing (which were repetitive) and direct travel, but were incapable of reliably detecting random travel that, by definition, lacked periodicity. The protocol was reactive and cumbersome, and the constant surveillance impinged on patient privacy. Parameters such as speed and directional changes, and path length were almost impossible to estimate reliably using observational methods. In our research, we were able to address the above-mentioned problem arising from the visual method by presenting a tool to analyze wandering behavior and categorize it automatically into Martino-Saltzman’s wandering pattern schema of direct travel, random, lapping or pacing. Prior research that has dichotomized movement data into direct and random patterns has resulted in little advancement in the understanding of relatively infrequent lapping and pacing behaviors. Increased availability of this new tool should yield a deeper understanding of these wandering subtypes and their relationship to clinical measures such as the MMSE.We have also contrasted it with Fractal Dimension; a spatial measure of randomness in a person’s movement-path that is moderately correlated with Mini Mental State Exam scores, dementia diagnosis, and fall likelihood PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5LZWFybnM8L0F1dGhvcj48WWVhcj4yMDEyPC9ZZWFyPjxS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ADDIN EN.CITE PEVuZE5vdGU+PENpdGU+PEF1dGhvcj5LZWFybnM8L0F1dGhvcj48WWVhcj4yMDEyPC9ZZWFyPjxS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ADDIN EN.CITE.DATA [16, 17, 99] and determined that the Fractal Dimension is strongly and positively associated with random and pacing travel but not lapping. Lapping, although a relatively infrequent event was paradoxically more likely to be present in subjects with no diagnosis of dementia. Revisiting the research objectives In Chapter-1 and Chapter-2, the author covered the basic concepts and advancement of research in the field of wandering technologies. In these chapters, the author defined the wandering from the prospect of stake-holder. This was very important to give direction for our research. The author did a thorough review of literature in Chapter-2 and studied many technologies and the way it has been implemented for wandering detection. In Chapter 3, the author discussed a novel algorithm for wandering pattern detection based on the Martino-Saltzman classification of pattern. This was a new approach, where a grid-based hybrid algorithm can be applicable for indoor as well outdoor navigation data. At the end of this chapter, the author also proposed a framework which can be used as an end-to-end solution for the wandering management which aligns with the first objective to develop a ubiquitous wandering management tool.In Chapter-4, the author validated the algorithm on the real-world dataset. The author also contrasted the method with previous work. In one example, the author compared the method with an alternative approach used by William D. Kearns and team for quantifying the wandering measures. In the result section, it was shown how the wandering typology contrasts with this method and what new insight we can get. MMSE score, which is used as the measure of cognition state of a person correlated significantly with the random pattern. However, quantization of these patterns from the navigation data is particularly important to establish the cognitive state of a person. Caregivers and family members can use this information to mitigate the problem associated with wandering. A statistical data mining approach to unveil hidden relation between wandering and factors such as MMSE formed the third objective to find the correlation of wandering with a cognitive test. It was also noticed that from coded observations, a variety of metrics can be generated for individual travel patterns and for wandering overall. Both frequency and duration of wandering episodes can be derived from navigation data, and both are important to consider, as each parameter may yield a somewhat different picture of wandering. Both frequency (as a sum, mean, or rate of wandering episodes, by pattern or collectively, per unit of time) and duration (as a sum, mean, or percentage of time spent wandering, by pattern or collectively) can be calculated for an observation period or across multiple observation periods, displayed graphically in relation to time of day, and used in statistical procedures. The grid-based approach allowed to realize the forth objective to study the fine-grained features in wandering.In Chapter-5, the author did a longitudinal study on the dataset of 10 subjects, which was the second objective to study wandering on long-term dataset. The data collected for 1 years for 10 subjects staying in an assisted living facility were able to uncover some consistent trend in the features associated with the navigation data. The subjects which showed consistent declining trend were found to be having a very low MMSE score. This was able to give useful insight into the progression of dementia. Combining everything it can be used to manage wandering in institution or home setup, which forms the fifth and final objective of this research.Limitation of our researchA major limitation of this study is that the study has been conducted on 25 subjects for one-month data and 10 subjects on one-year data. This is because of the unavailability of such data set. For our knowledge this has been the first instance of reporting the pattern at this scale. Despite these limitations, this study has demonstrated that statistical features of wandering can be automatically studied by using technology such as UWB or other electronics techniques. We hope that the developed analytic approach will stimulate further research on wandering, particularly with regard to the fine features of wandering that maximize the travel independence of cognitively impaired older adults without compromising their safety.Our interest was to differentiate movement variability in ALF residents into the Martino Saltzman categorical system and then relate the proportional time and distance in category to clinical and empirical measures of cognitive impairment and observed wandering. A secondary goal was to evaluate the categorical system against Kearns’ use of Fractal Dimension as a means to illuminate the relationships among the two systems of measurement (one categorical [Martino Saltzman], the other continuous [Fractal D]). With progressive refinement of indoor Real Time Location Sensing systems and improved algorithmic approaches to evaluating path variability in persons with cognitive impairment, we should theoretically be capable of resolving finer gradations of cognitive impairment linked to the presence of MCI. Although the presence of a clinical diagnosis of dementia was found in the records of a significant proportion of the sample, it was evident from the Mini Mental State Exam that a number of subjects had scores indicating significant cognitive impairment with no clinical diagnosis of dementia. Future research in this area should include a more precise evaluation of the current clinical status of the subjects.Adoption of technology for wandering managementTechnology can play an important role in the management of wandering in dementia. But the major concern remains with the adoption of these technologies by various stakeholders such as family, caretakers or the person having dementia. In a study, Kearns ADDIN EN.CITE <EndNote><Cite><Author>Kearns</Author><Year>2007</Year><RecNum>547</RecNum><DisplayText>[100]</DisplayText><record><rec-number>547</rec-number><foreign-keys><key app="EN" db-id="5afv0p2dswettmets25pe2pg2fzt29999efv" timestamp="1534783756">547</key></foreign-keys><ref-type name="Journal Article">17</ref-type><contributors><authors><author>Kearns, William D</author><author>Rosenberg, Deborah</author><author>West, Lisa</author><author>Applegarth, Shawn</author></authors></contributors><titles><title>Attitudes and expectations of technologies to manage wandering behavior in persons with dementia</title><secondary-title>Gerontechnology</secondary-title></titles><periodical><full-title>Gerontechnology</full-title></periodical><pages>89</pages><volume>6</volume><number>2</number><dates><year>2007</year></dates><urls></urls></record></Cite></EndNote>[100] obtained the opinion on the potential effectiveness of existing technologies for managing wandering. Participants were divided into 6 groups based on the roles they play in dementia care. A consensus-based analysis was done, and these are the outcome of the study. 1. Almost all the group wanted a flexible system that would allow for a normal daily life without infringing on the user’s privacy. An inconspicuous device resembling a common necklace pendant (for females) or a watch (for males) was thought preferable as it does not stigmatize the person as having an illness or disability. 2. The cost was one of the main factors which was considered important while using such technologies as most of these wandering management devices are not covered under insurance. 3. Inside the home, motion sensor was preferred over the camera-based system and GPS for the outdoor monitoring. 4. All the groups wanted the technology to be less restrictive in nature and it should help the person with dementia to live life with dignity.Our sensor-based wandering detection and management system has been developed keeping the stakeholders’ concern in mind. It is very cost effective and infringes minimum with the user’s privacy. It is very easy to use as the sensor is in the form of a watch and it is not restrictive in nature. Future WorkCollection of larger dataset for longitudinal studyAs discussed in the previous section, due to the very high cost and infrastructure requirement for the data collection, we were not able to collect a very large dataset. Since now we have our infrastructure ready in two ALFs in Tampa, Florida, we are in the process of collecting more such data for our research. This will be further used to validate our algorithm and to get new insight in terms of more reliable trends and features in wandering.Activity recognition approach for contextual wandering analysisDue to the low resolution of data, we were only limited to identification of wandering pattern in an episode. The activity performed in between the intervening episodes (no walking phase) as shown in REF _Ref504684765 \h Figure 6.1 could not be identified. This could have been used to find the contextual relationship about the activity in wandering episodes. Type and duration of the activities performed during these periods can be important to understand the intent of wandering in dementia patients.Figure 6.1: Phases in wanderingPersonalized User wandering detectionDespite the general pattern of wandering such as random, lapping or pacing, it can have different spatiotemporal characteristics for a different individual. It can be in the same environmental condition or a particular person in different environmental condition. Wandering may be manifested differently in different condition. Therefore, a personalized wandering detection algorithm tailored to specific individual and condition can be the ultimate solution for the wandering management. AppendixList of publication ADDIN EN.REFLIST 1.Ashish Kumar, Chiew Tong Lau, Syin Chan, Maode ma, William Kearns. Trend Analysis in the Trajectory of the Dementia Patients. In Proc. of ICSEC 2017.2.Ashish Kumar, Syin Chan, Chiew Tong Lau, Maode ma. PEAR: An app for Person-Centred Care of Dementia Patients. In Proc. of ICAA 20173.Ashish Kumar, Maode ma, Chiew Tong Lau, Syin Chan. A Framework of Real-time Wandering Management for Person with Dementia. In Proc. of ICCMS 20174.Ashish Kumar, Chiew Tong Lau, Syin Chan, Maode ma, William Kearns. A Longitudinal Study of the Navigation Patterns of Dementia Patients and their Relationship to MMSE. In Proc. of IAGG 20175.Ashish Kumar, Chiew Tong Lau, Syin Chan, Maode ma, William Kearns. A Unified Grid-based Wandering Pattern Detection Algorithm. In Proc. of EMBC 2016W.D. Kearns, J.L. Fozard, A. Kumar. Longitudinal analysis of ALF resident wandering using real-time location services: Results of a yearlong study. Gerontechnology 2017.Under ReviewAshish Kumar, William Kearns, Syin Chan, Chiew Tong Lau, Maode ma. A Data Analytic Study of Wandering Trajectories of Assisted Living Facility Residents with Dementia. Journal of Biomedical and Health Informatics (J-BHI) References ADDIN EN.REFLIST 1.Wentzel, C., et al., Progression of impairment in patients with vascular cognitive impairment without dementia. Neurology, 2001. 57(4): p. 714-716.2.Rodrigues, R., M. Huber, and G. Lamura, Facts and figures on healthy ageing and long-term care. Vienna: European Centre for Social Welfare Policy and Research, 2012.3.Suzman, R., et al., Health in an ageing world—what do we know? The Lancet, 2015. 385(9967): p. 484-486.4.Silverstein, N., G. Flaherty, and T. Tobin, Comprehensive review of the literature on wandering behavior. Dementia and wandering behavior: concern for the lost elder. New York: Springer, 2002: p. 28-51.5.Pollock, B.G., et al., A double-blind comparison of citalopram and risperidone for the treatment of behavioral and psychotic symptoms associated with dementia. The American Journal of Geriatric Psychiatry, 2007. 15(11): p. 942-952.6.Hope, T., et al., Wandering in dementia: a longitudinal study. International Psychogeriatrics, 2001. 13(02): p. 137-147.7.Martino-Saltzman, D., et al., Travel behavior of nursing home residents perceived as wanderers and nonwanderers. The Gerontologist, 1991. 31(5): p. 666-672.8.Algase, D.L., E.R. Beattie, and B. Therrien, Impact of cognitive impairment on wandering behavior. Western Journal of Nursing Research, 2001. 23(3): p. 283-295.9.Holden, J.E. and B. Therrien, Cue familiarity reduces spatial disorientation following hippocampal damage. Nursing research, 1993. 42(6): p. 338-343.10.Neistein, S. and A.P. Siegal, Agitation, wandering, pacing, restlessness, and repetitive mannerisms. International Psychogeriatrics, 1997. 8(S3): p. 399-402.11.Algase, D.L., Wandering in dementia. Annual review of nursing research, 1999. 17(1): p. 185-217.12.Herdman, T.H., Nursing diagnoses 2012-14: Definitions and classification. 2011: John Wiley & Sons.13.Algase, D.L., et al., The Algase Wandering Scale: initial psychometrics of a new caregiver reporting tool. American Journal of Alzheimer's Disease & Other Dementias?, 2001. 16(3): p. 141-152.14.Algase, D.L., et al., Estimates of stability of daily wandering behavior among cognitively impaired long-term care residents. Nursing Research, 1997. 46(3): p. 172-178.15.Kearns, W.D., V. Nams, and J.L. Fozard, Tortuosity in movement paths is related to cognitive impairment. Methods Inf Med, 2010. 49(6): p. 592-598.16.Kearns, W.D., J.L. Fozard, and V.O. Nams, Movement path tortuosity in free ambulation: Relationships to age and brain disease. IEEE Journal of Biomedical and Health Informatics, 2016. 17.Kearns, W.D., et al., Path Tortuosity in Everyday Movements of Elderly Persons Increases Fall Prediction Beyond Knowledge of Fall History, Medication Use, and Standardized Gait and Balance Assessments. Journal of the American Medical Directors Association, 2012. 13(7): p. 665.e7-665.e13.18.Mulvenna, M., et al. Designing & evaluating a cognitive prosthetic for people with mild dementia. in Proceedings of the 28th Annual European Conference on Cognitive Ergonomics. 2010. ACM.19.Lin, Q., et al. Detecting wandering behavior based on GPS traces for elders with dementia. in Control Automation Robotics & Vision (ICARCV), 2012 12th International Conference on. 2012. IEEE.20.Shimizu, K., K. Kawamura, and K. Yamamoto. Location system for dementia wandering. in Engineering in Medicine and Biology Society, 2000. Proceedings of the 22nd Annual International Conference of the IEEE. 2000. IEEE.21.Ota, K., et al. Elderly-care motion sensor using UWB-IR. in Sensors Applications Symposium (SAS), 2011 IEEE. 2011. IEEE.22.Kearns, W. and J. Fozard, Evaluation of wandering by residents in an assisted living facility (ALF) using ultra wideband radio RTLS. 2009.23.Kearns, W.D., et al., Ultra wideband radio: A novel method for measuring wandering in persons with dementia. Gerontechnology, 2008. 7(1): p. 48.24.Kearns, W.D., et al., Wireless telesurveillance system for detecting dementia. Gerontechnology, 2011. 10(2).25.Lim, C.-H., et al., A real-time indoor WiFi localization system utilizing smart antennas. IEEE Transactions on Consumer Electronics, 2007. 53(2).26.Nugent, C. and J. Augusto. A system for activity monitoring and patient tracking in a smart hospital. in Smart Homes and Beyond: ICOST 2006: 4th International Conference on Smart Homes and Health Telematics. 2006. IOS Press.27.Miskelly, F., Electronic tracking of patients with dementia and wandering using mobile phone technology. Age and ageing, 2005. 34(5): p. 497-498.28.Smith, M., et al., Dementia care in assisted living: Needs and challenges. Issues in mental health nursing, 2008. 29(8): p. 817-838.29.Katz, S. and C.A. Akpom, A measure of primary sociobiological functions. International journal of health services, 1976. 6(3): p. 493-508.30.Alwan, M., et al. A smart and passive floor-vibration based fall detector for elderly. in Information and Communication Technologies. 2006.31.Sixsmith, A. and N. Johnson, A smart sensor to detect the falls of the elderly. IEEE Pervasive computing, 2004(2): p. 42-47.32.Bharucha, A.J., et al., Intelligent assistive technology applications to dementia care: current capabilities, limitations, and future challenges. The American journal of geriatric psychiatry, 2009. 17(2): p. 88-104.33.Rowe, M., S. Lane, and C. Phipps, CareWatch: a home monitoring system for use in homes of persons with cognitive impairment. Topics in geriatric rehabilitation, 2007. 23(1): p. 3.34.Lee, T. and A. Mihailidis, An intelligent emergency response system: preliminary development and testing of automated fall detection. Journal of telemedicine and telecare, 2005. 11(4): p. 194-198.35.Nait-Charif, H. and S.J. McKenna. Activity summarisation and fall detection in a supportive home environment. in Pattern Recognition, 2004. ICPR 2004. Proceedings of the 17th International Conference on. 2004. IEEE.36.Doughty, K., et al. DIANA-a telecare system for supporting dementia sufferers in the community. in Engineering in Medicine and Biology Society, 1998. Proceedings of the 20th Annual International Conference of the IEEE. 1998. IEEE.37.Jit, B., et al., A system for activity monitoring and patient tracking in a smart hospital. Proceedings of the International Conference on Smart Homes and Health Telematics, ICOST 2006, 2006: p. 196-203.38.Masuda, Y., et al. Unconstrained monitoring of prevention of wandering the elderly. in Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2002.39.Ota, K., et al. Elderly-care motion sensor using UWB-IR. in SAS 2011 - IEEE Sensors Applications Symposium, Proceedings. 2011.40.Rowe, M., S. Lane, and C. Phipps, CareWatch: A home monitoring system for use in homes of persons with cognitive impairment. Topics in Geriatric Rehabilitation, 2007. 23(1): p. 3-8.41.Martino-Saltzman, D., et al., Travel behavior of nursing home residents perceived as wanderers and nonwanderers. Gerontologist, 1991. 31(5): p. 666-672.42.Algase, D.L., et al., Biomechanical activity devices to index wandering behavior in dementia. American Journal of Alzheimer's Disease and other Dementias, 2003. 18(2): p. 85-92.43.Kearns, W. and J.L. Fozard, Evaluation of Wandering by Residents in an Assisted Living Facility (ALF) using Ultra Wideband Radio RTLS. J Nutrition Health Aging, 2009. 13.44.Kearns, W.D., et al., Ultra wideband radio: A novel method for measuring wandering in persons with dementia. Gerontechnology, 2008. 7(1): p. 48-57.45.Kearns, W.D., et al., Wireless telesurveillance system for detecting dementia. Gerontechnology, 2011. 10(2): p. 90-102.46.Kearns, W.D., V.O. Nams, and J.L. Fozard, Tortuosity in movement paths is related to cognitive impairment wireless fractal Estimation in assisted living facility residents. Methods of Information in Medicine, 2010. 49(6): p. 592-598.47.Nams, V.O. and M. Bourgeois, Fractal analysis measures habitat use at different spatial scales: An example with American marten. Canadian Journal of Zoology, 2004. 82(11): p. 1738-1747.48.Taub, D.M., et al., The Escort system: A safety monitor for people living with Alzheimer's disease. IEEE Pervasive Computing, 2011. 10(2): p. 68-76.49.Vuong, N.K., S. Chan, and C.T. Lau, Feasibility study of a real-time wandering detection algorithm for dementia patients. Proc Mobile Health 2011, 2011.50.Lin, Q., et al. Detecting wandering behavior based on GPS traces for elders with dementia. in 2012 12th International Conference on Control, Automation, Robotics and Vision, ICARCV 2012. 2012.51.Liao, L., D. Fox, and H. Kautz. Learning and inferring transportation routines. in Proceedings of the National Conference on Artificial Intelligence. 2004.52.Chang, Y.J., Anomaly detection for travelling individuals with cognitive impairments. SIGACCESS Newsletter, 2010. 97: p. 25-32.53.Chang, Y.J. and T.Y. Wang, Mobile location-based social networking in supported employment for people with cognitive impairments. Cybernetics and Systems, 2010. 41(3): p. 245-261.54.Lim, C.H., et al., A real-time indoor WiFi localization system utilizing smart antennas. IEEE Transactions on Consumer Electronics, 2007. 53(2): p. 618-622.55.Chen, D., A.J. Bharucha, and H.D. Wactlar. Intelligent video monitoring to improve safety of older persons. in Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2007.56.Vuong, N.K., et al. A predictive location-aware algorithm for dementia care. in Proceedings of the International Symposium on Consumer Electronics, ISCE. 2011.57.Sposaro, F., J. Danielson, and G. Tyson. IWander: An Android application for dementia patients. in 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC'10. 2010.58.Wan, J., et al., OutCare: Supporting dementia patients in outdoor scenarios, in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 2010. p. 365-374.59.Wan, J., et al. Orange alerts: Lessons from an outdoor case study. in 2011 5th International Conference on Pervasive Computing Technologies for Healthcare and Workshops, PervasiveHealth 2011. 2011.60.Hoey, J., X. Yang, and J. Favela, Decision theoretic, context aware safety assistance for persons who wander. Proceedings of the the 7th International Workshop on Ubiquitous Health and Wellness (UbiHealth 2012), 0000.61.Hoey, J., et al. LaCasa: Location and context-aware safety assistant. in 2012 6th International Conference on Pervasive Computing Technologies for Healthcare and Workshops, PervasiveHealth 2012. 2012.62.Ogawa, H., et al. A mobile phone-based safety support system for wandering elderly persons. in Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2004.63.Rodríguez, M.D., et al. An ontological representation model to tailor ambient assisted interventions for wandering. in AAAI Fall Symposium - Technical Report. 2012.64.Miskelly, F., Electronic tracking of patients with dementia and wandering using mobile phone technology [1]. Age and Ageing, 2005. 34(5): p. 497-499.65.Shimizu, K., K. Kawamura, and K. Yamamoto. Location system for dementia wandering. in Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings. 2000.66.Calvo-Palomino, R., et al., Outdoors monitoring of elderly people assisted by compass, GPS and mobile social network, in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 2009. p. 808-811.67.Lin, C.C., et al., Wireless health care service system for elderly with dementia. IEEE Transactions on Information Technology in Biomedicine, 2006. 10(4): p. 696-704.68.Mulvenna, M., et al. Designing & evaluating a cognitive prosthetic for people with mild dementia. in ECCE 2010 - European Conference on Cognitive Ergonomics 2010: The 28th Annual Conference of the European Association of Cognitive Ergonomics. 2010.69.Wherify Wireless, 0000.70.Parnes, P.B., GPS Technology and Alzheimer's Disease: Novel Use for An Existing Technology, 0000.71.Miskelly, F., A novel system of electronic tagging in patients with dementia and wandering. Age and Ageing, 2004. 33(3): p. 304-306.72.Blackburn, P., Freedom to wander. Nursing times, 1988. 84(49): p. 54-55.73.APDM. APDM Wearable Technologies. 2017 [cited 2017 22/09/2017]; Available from: .74.Roetenberg, D., H. Luinge, and P. Slycke, Xsens MVN: full 6DOF human motion tracking using miniature inertial sensors. Xsens Motion Technologies BV, Tech. Rep, 2009.75.Scalisi, J.F. and L.E. Hecox, Communication system and method including communication channel mobility. 2007, Google Patents.76.Kim, K.-J., et al. Dementia wandering detection and activity recognition algorithm using tri-axial accelerometer sensors. in Ubiquitous Information Technologies & Applications, 2009. ICUT'09. Proceedings of the 4th International Conference on. 2009. IEEE.77.Vuong, N., et al. A mobile-health application to detect wandering patterns of elderly people in home environment. in Engineering in Medicine and Biology Society (EMBC), 2013 35th Annual International Conference of the IEEE. 2013. IEEE.78.Vuong, N., S. Chan, and C. Lau, Automated detection of wandering patterns in people with dementia. Gerontechnology, 2014. 12(3): p. 127-147.79.Kearns, W.D., J.L. Fozard, and V.O. Nams, Movement Path Tortuosity in Free Ambulation: Relationships to Age and Brain Disease. IEEE Journal of Biomedical and Health Informatics, 2016.80.Park, S. and S. Jayaraman, Enhancing the quality of life through wearable technology. IEEE Engineering in medicine and biology magazine, 2003. 22(3): p. 41-48.81.Wu, J.-H., S.-C. Wang, and L.-M. Lin, Mobile computing acceptance factors in the healthcare industry: A structural equation model. International journal of medical informatics, 2007. 76(1): p. 66-77.82.Takacs, J., et al., Validation of the Fitbit One activity monitor device during treadmill walking. Journal of Science and Medicine in Sport, 2014. 17(5): p. 496-500.83.Christensen, K., et al., Ageing populations: the challenges ahead. The lancet, 2009. 374(9696): p. 1196-1208.84.Finkel, S.I., et al., Behavioral and psychological signs and symptoms of dementia: a consensus statement on current knowledge and implications for research and treatment. International Psychogeriatrics, 1997. 8(S3): p. 497-500.85.Nelson, A. and D.L. Algase, Evidence-based protocols for managing wandering behaviors. 2007: Springer Publishing Company.86.Algase, D., et al., Mapping the maze of terms and definitions in dementia-related wandering. Aging & mental health, 2007. 11(6): p. 686-698.87.Kearns, W.D., et al., Wireless telesurveillance system for detecting dementia. Gerontechnology, 2011. 10(2): p. 90-102. doi: 10.4017/gt. 2011.10. 2.004. 00.88.Sposaro, F., J. Danielson, and G. Tyson. iWander: An Android application for dementia patients. in Engineering in Medicine and Biology Society (EMBC), 2010 Annual International Conference of the IEEE. 2010. IEEE.89.Hachour, O., Path planning of Autonomous Mobile robot. International journal of systems applications, engineering & development, 2008. 2(4): p. 178-190.90.Kearns, W., et al., Wireless telesurveillance system for detecting dementia. Gerontechnology, 2011. 10(2): p. 90-102.91.Zheng, Y., Trajectory data mining: an overview. ACM Transactions on Intelligent Systems and Technology (TIST), 2015. 6(3): p. 29.92.Folstein, M.F., S.E. Folstein, and P.R. McHugh, “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. Journal of psychiatric research, 1975. 12(3): p. 189-198.93.Algase, D.L., et al., The interrelatedness of wandering and wayfinding in a community sample of persons with dementia. Dementia and geriatric cognitive disorders, 2004. 17(3): p. 231-239.94.Kearns, W.D., et al., Location aware smart watch to support aging in place. Gerontechnology, 2014. 13(2).95.Vivago develops smart safety and wellbeing solutions for preventive care. 2017.96.Offender Monitoring | Numerex IoT Solutions. 2017.97.Folstein, S., Key Papers in Geriatric Psychiatry. Mini-Mental State: A Practical Method for Grading the Cognitive State of Patients for the Clinician. M. Folstein, S. Folstein and P. McHugh, Journal of Psychiatric Research (1975) 12, 189_198. International Journal of Geriatric Psychiatry, 1998. 13: p. 285-294.98.Algase, D.L., et al., The interrelatedness of wandering and wayfinding in a community sample of persons with dementia. Dementia & Geriatric Cognitive Disorders, 2004. 17(3): p. 231-9.99.Kearns, W.D., J.L. Fozard, and V.O. Nams, Movement path tortuosity in free ambulation: relationships to age and brain disease. IEEE journal of biomedical and health informatics, 2017. 21(2): p. 539-548.100.Kearns, W.D., et al., Attitudes and expectations of technologies to manage wandering behavior in persons with dementia. Gerontechnology, 2007. 6(2): p. 89. ................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- technological change in business
- technological change examples
- technological advancements in accounting
- technological advancements in 2010
- what is technological change
- technological advancements in history
- the effects of technological changes
- major technological advancements
- technological change definition
- 10 greatest technological advances
- technological advances in healthcare
- technological changes in business