Professional Report - University Of Maryland



Indoor & Outdoor Location of

Firefighters & Safety Personnel

An Integrated Approach

Final Report Submitted to

Maryland Fire and Rescue Institute

Submitted by

|Dr. Neil Goldsman | |Dr. Gilmer Blankenship | |

|(301)405-3648 | |(240)790-0620 | |

|neil@eng.umd.edu | |gilmer@ | |

| | |

| | |

|April 14, 2006 | |

Contents

Section I. Abstract 1

Section II. Final Report 2

a. Introduction 2

b. Indoor / Outdoor Tracking Performance Trials 4

c. Detailed Technical Description and Approach 6

c.1. Integrated Positioning 6

c.2. Elements of the Existing Sentinel System 8

c.3. 3D Graphical User Interface 11

d. Miniaturization of Prototype 14

d.1. INU Improvements 17

e. Node Controller 20

f. Future Work 21

Figures

Figure 1. Fire Sentinel Beacon has 2 pieces, the INU is worn on the firefighters belt. The INU is connected to the radio via blue tooth. The radio can be connected in any convenient location so as not to interfere with movement or other gear. 2

Figure 2 (Left) Trajectory produced by the Sentinel INU of a user walking on the third floor of the AV Williams Bldg. (Right) The actual path the user walked inside the building. The journey lasted about 5 minutes. 5

Figure 3 Shows a path traversed outdoors, around a relatively large building by a user of the prototype Sentinel Beacon. The path is 2/5 of a mile. No GPS is used. 6

Figure 4 Inertial Navigation Unit 9

Figure 5 Transceiver module and RSSI display 9

Figure 6 Motherboard of prototype sensor interface. 10

Figure 7 GPS System designed by TRX 10

Figure 13. System Block Diagram 11

Figure 8 A part of the path shown previously: (Left) Path plotted (in real-time) on a floor plan using prototype map-matching Virtual Reality software. (Right) 3 dimensional VR view of the unit moving on the floor plan. 13

Figure 9 Building internal structure constructed by JAVA3D map-building software for the path shown earlier using data provided by the INU. 14

Figure 10. First Generation INU Block Diagram 17

Figure 11. Second Generation INU Block Diagram 17

Figure 12. Second Generation INU Boards a) INU main board is worn vertically with back against firefighter’s belt. b) INU secondary board inserts into the main board at silk screened rectangle and allows for tracking of firefighters in non-vertical orientations such as crawling. 19

Figure 14 Liteye HUD mounted on a helmet. The viewer is “transparent” enabling the user to see objects ahead. 23

Abstract

In work funded by the Department of Homeland Security (DHS), the Maryland Industrial Partnerships Program (MIPS) and the Maryland Fire and Rescue Institute (MFRI), an innovative technology prototype for tracking individuals moving inside buildings and structures (and outside as well) has been developed. Specifically, the Sentinel System is designed to locate and track firefighters and other first responders indoors and outdoors using inertial and other measurements, including GPS signals, if available. The system includes a Sentinel Beacon and a “Command Station” (a ruggedized PC) that receives the data transmitted from the beacons and processes it to produce location and track information for the incident commander. The Sentinel Beacon system is designed to work as a “mesh network” so that a beacon that is out of range of the command station can have its data relayed by other beacons.

We have been able to leverage the ongoing work for DHS and MIPS to allow us to accomplish the results presented here. The funding from the University of Maryland Fire and Rescue Institute has been used to support the miniaturization of the prototype hardware.

Final Report

[pic]

Figure 1. The second generation Fire Sentinel Beacon has 2 pieces, the INU is worn on the firefighters belt. The INU is connected to the radio via blue tooth. The radio can be connected in any convenient location so as not to interfere with movement or other gear.

Our previous work established a strong foundation for the project by developing and implementing a prototype integrated positioning technology for indoor and outdoor tracking. The development of the prototype system has been supported by the US Department of Homeland Security (DHS)/Technology Support Working Group (TSWG) and the Maryland Industrial Partnerships Program (MIPS). We have been able to leverage the ongoing work for DHS and MIPS to allow us to accomplish the results presented here. The funding from the University of Maryland Fire and Rescue Institute (MFRI) has been used to support the miniaturization of the prototype hardware.

1 Introduction

The ability to accurately track first responders at incident sites both inside and outside structures is a top priority for safety and command and control. Our Sentinel System is designed to enable the Incident Commander to locate and track responders inside a building, or in the vicinity of an incident. The Sentinel System consists of the following elements:

■ A compact, body worn unit (a Personal Locator Beacon – PLB) that computes the responder’s location in absolute coordinates and communicates it locally to the Incident Commander. The capability to relay data from out of range nodes is also provided by PLB units;

■ A Base Station that enables the Incident Commander to see the locations of all the responders in the squads (at the scene, both inside and outside), together with other information such as responder health, equipment condition, environmental data, building plan, assets, etc.; and

■ A Communications Network that links all responders and the Commander(s).

The initial product will compute location and measure basic health status of the responder (body temperature, pulse rate, net movement, impact alarm, etc.) and transmit location, health status, and alarms to the base station. It will transmit an alarm when the responder’s status leaves a nominal range, or when the responder activates the alarm manually.

The base station will show the Incident Commander where the responders are as mapped on a 3-D “floor plan” displayed using a graphical user interface. The personnel locations and tracks will be shown on a building plan or, if a building plan is not available, a map of the building will be developed implicitly from the historical path data.

An important feature of our system is that, all of the mobile units form a mesh network so if a responder goes out of range of the base station transceiver, as long another mobile unit is in range of the responder and base station, the information can be forwarded to the base station, effectively extending the range using multiple hops to get out of the building. Finally, the environmental data collected by the sensors in the unit can be used to create a “map” of the environmental conditions inside the structure. No other system has this collection of broad and rich capabilities.

With our system, if a responder goes down

the Commander will know where they are and how they got there.

While the responders will go after a comrade with that knowledge, it may still difficult to find their way to the downed comrade if, for example, the room is filled with smoke, so our mobile units will have LED indicators to indicate direction and distance to a downed responder. Future designs will include a heads up display to aid in locating responders. The display will include a map of the individual’s path through the structure and other mission specific critical data.

2 Indoor / Outdoor Tracking Performance Trials

[pic][pic]

Figure 2 (Left) Trajectory produced by the Sentinel INU of a user walking on the third floor of the AV Williams Bldg.

(Right) The actual path the user walked inside the building. The journey lasted about 5 minutes.

In the figures that follow we summarize evaluations of the performance of the current Sentinel System in tracking individuals inside and outside a building. The tests were conducted at the AV Williams Building[1] at the University of Maryland in College Park in February 2006. This is a four story building with a complex internal structure. It houses the Electrical Engineering and Computer Science Departments, and contains a large amount of electrical equipment – a potential source of interference with the Sentinel System.

The first trial (Figure 1) shows the track of an individual walking in the hallways of the third floor of the building. The track was taken in a continuous experiment that lasted about 5 minutes. Note the straight lines of the path corresponding to the long corridors in the building, and the sharp 90o changes in the track as the individual turns each corner. Note also the accuracy of the unit as the individual moves back along the same corridors. Since the return paths overlap the outward bound paths, we conclude that the location error is less than the width of the corridor (6 feet). As the data trajectory indicates, the INU sends more data when it perceives a turn in the path.

A video demonstrating the indoor tracking capabilities of the Sentinel System is available for review – specifically a recording of the experiment shown in Figure 1. This video can be viewed by downloading a Windows Media file from the TRX Systems website:



Login as “MFRI” with password “trackme”. Download the file named “TRX Tracking Demo” to your computer and view it using the Windows Media Player.[2]

The trial displayed in Figure 2 is a path of about 2/5 mile traversed outside the building. The individual stayed as close to the walls of the building as possible during the journey (to provide a reference for accuracy). The trajectory is overlaid on a satellite image of the building from the Google Earth web site. Although the experiment was outside, we did not use GPS data – only the Sentinel INU. If typical GPS accuracy is 2-10 meters, then the location results from the Sentinel unit are better than GPS.

[pic]

Figure 3 Shows a path traversed outdoors, around a relatively large building

by a user of the prototype Sentinel Beacon. The path is 2/5 of a mile. No GPS is used.

3 Detailed Technical Description and Approach

We call our approach Integrated Positioning. Integrated positioning combines several technologies to obtain accurate position estimates. The integration of the different technologies is seamless to the user and provides more robust, accurate position estimates than a single technique used alone. The final Personal Locator Beacon (PLB) unit worn by a first responder will contain all of the integrated positioning subsystems and will be approximately the size of a cell phone. The information obtained by the mobile PLB, is then transmitted to the base station. In addition to communicating with the base station, the mobile PLB also supports communication of positioning and condition data to other responders at the incident.

1 Integrated Positioning

The positioning technologies used in Integrated Positioning are summarized below:

• Inertial Navigation Unit (INU): Contains a MEMS-accelerometer, MEMS-compass and MEMS-gyroscope to track an individual’s motion. We have already developed a first generation of the hardware, most of the signal processing software, and a functioning prototype. The accelerometers are also used to determine if a responder is not moving or undergoes a rapid acceleration, e.g., falls through a floor or gets knocked over.

The Sentinel INU is targeted toward monitoring the position of personnel and is capable of accurate locations over a longer period of time compared to conventional INU’s. The data processing methods are targeted specifically toward recognizing distinctive characteristics in human gait and to take advantage the gait properties and their relation to movement. By monitoring and analyzing specific characteristics, the INU’s processor uses specific filtering techniques to compare the recorded motions with a library of predefined gait motions.

• Global Positioning (GPS): A small unit based on the latest GPS receiver chips that can track at very low signal to noise ratios (SNR). We have designed the hardware and software that enables the realization of this chip in a GPS receiver. A prototype of this subsystem about the size of a postage stamp has already been developed and integrated into the Sentinel Beacon.

• Received Signal Strength Indication (RSSI): Each base station and PLB will have the capability to measure the signal strength of received radio signals. The weaker the signal, the further the responder is from the signal source. This method is also used to determine distance between responders in the network. RSSI becomes very accurate when the transmit/receive pair are in close proximity to each other, for example, in a room or large open area. RSSI can be used as a backup to locate downed responders in a situation where vision is impaired such as in a smoke filled room. A prototype for hardware and software has been developed.

• Mesh Network: The mobile nodes in our sentinel system form a mesh network. By forming a mesh network data is allowed to be routed (“hopped”) though other nodes if the nodes cannot form a direct link with the base station. The ability to hop extends the range of the system. Each member of the team is thus networked to each other, and each node in the network can know the location and other pertinent data of the other nodes. A working system has been tested in prototype form.

• Active Radar: This location method uses extremely accurate GHz clocks and electromagnetic ultra-wide-band pulses for positioning. This method is very similar to police and aircraft radar. However, it uses ultra-wide band technology and transceiver pairs as opposed to the passive reflection of radio waves employed by existing systems. Theory, algorithms, and proof of concept hardware have been developed.

2 Elements of the Existing Sentinel System

The positioning technology uses multiple sensors, subsystems, and algorithms to determine the location of the tracked personnel. The data from each subsystem is fused onboard the Sentinel Beacon and in the Sentinel Command Station. Using data fusion and complex algorithms, the integrated positioning software is able to locate personnel both indoors and out with excellent accuracy. The subsystems are integrated into a small motherboard that coordinates their operation.

The subsystems used in the integrated positioning system are summarized below:

|[pic] |

|Figure 4 Inertial Navigation Unit |

Inertial Navigation Unit (INU): The INU subsystem contains a MEMS accelerometer, compass and gyroscope. The information from these micro-sensors is combined using novel signal processing algorithms, programmed into microprocessors, to track the path taken by an individual on foot. We use this technique inside buildings and structures where GPS is not effective.

Received Signal Strength Indication (RSSI): This subsystem uses the radio signal strength of the beacon transceiver to measure the distance from an individual to the Command Station. This method is useful for homing in on an individual during a search and rescue operation. Figure 5 shows a prototype of the transceiver equipped with RSSI capabilities and the base-station interface with the RSSI data. The circle indicates the distance of an individual from the Base Unit.

[pic]

Figure 5 Transceiver module and RSSI display

Motherboard: In Figure 6 we show the motherboard for the current Sentinel Beacon. This unit is currently being miniaturized under funding from DHS. The motherboard was designed to operate with a wide assortment of plug-and-play sensors. (TRX Systems developed the interface hardware and software that allowed these sensors to be plug-and-play.) This is reflected by the eight rows of connector pins, where various sensors plug into the motherboard, depending on the application desired by the user.

[pic]

Figure 6 Motherboard of prototype sensor interface.

|[pic] |

|Figure 7 GPS System designed by TRX |

We will continue to provide an option for adding a plug-and-play modular sensors, as desired by the user. The PIC motherboard microprocessor will be programmed to interpret the data obtained from these plug and play sensors, and incorporate the information into the node report that will be sent to the Base Station.

Global Positioning (GPS) Unit: This is a subunit that employs the latest single chip GPS receivers, which is shown in Figure 7. The GPS unit can usually provide tracking information outdoors. GPS is not effective for indoor tracking.

3 3D Graphical User Interface

[pic]

Figure 8. System Block Diagram

The Sentinel Base Station is a Panasonic Toughbook, a ruggedized laptop that runs standard Windows operating systems. The user interface can plot the trajectory of the personnel and archive the data. Future plans include enhancement the base station with algorithms which help to eliminate errors, and provide even more accurate values for the trajectories. If building floor plans or structure (tunnel) maps are available, the trajectories will be plotted on them for ease of observation and for corrections based on the structure.

The laptop can be connected to a Wide Area Network (WAN) modem for distribution of data to other surveillance locations. The Toughbook has an 80-pin reinforced port-replicator port which allows it to be connected to a preinstalled set of peripherals, such as a wireless modem. The Base Station Transceiver will be developed by TRX Systems. It will be a transceiver for receiving and interpreting information from the Sentinel PLBs and interfacing with the Command Station. It will receive the node reports transmitted by the PLBs, and have a microprocessor to control communications. The Base Station Transceiver is smaller than the size of a small cell phone.

The interface concept design is shown in Figure 8 in prototype form. The base station user interface will run on a ruggedized laptop. The interface will be capable of displaying current and past location and sensor information for all mobile units on a building floor plan – if available. It will have 3D and 2D viewing options with natural transitions among views, for example, click on a floor in the 3D move to change to 2D view.

A whole building wire-frame view of the building structure will be developed with the position of each responder indicated. The building will include a scale showing grid spaces of approximately 10 feet in every direction. Positions will be shown for all responders, if an alarm is detected the subject’s marker will be changed to red and enlarged or blink. Detailed information on sensor data, etc. can be seen by clicking on an individual marker. Historical path data can be shown if desired, for a selected individual, group, or for all mobile nodes.

The user interface will allow the incident commander to group individuals into teams (by task, by engine,…) that can be represented by different color or shape markers for ease of tracking.

We are experimenting with various 3D software packages for display of the data. A key is to provide a system of maximum effectiveness to field commanders – the GUI cannot be overly complex. It must make the information, especially alarms, very easy to understand in an operational setting. We will work with MFRI and DHS to achieve an effective design.

The Base Station software is critical for displaying the positions of responders for decisive analysis and action. To enhance the accuracy of indoor tracking, and to aid visualization, we shall develop algorithms and software to project the path of the first responder on pre-existing maps[3] or to create maps created based on the subject’s trajectory.

The key features of this task are: Map Matching; Map Building; and Outdoor Terrain Mapping.

[pic] [pic]

Figure 9 A part of the path shown previously:

(Left) Path plotted (in real-time) on a floor plan using prototype map-matching Virtual Reality software.

(Right) 3 dimensional VR view of the unit moving on the floor plan.

As shown in the prototype design in Figure 8, the software will plot the current location and maintain a database of previous positions to produce a “trail” of the path taken by the first responder, if desired. The historical data will also be used to provide an incident playback feature.

The program shown in Figure 8 was designed using the Virtual Reality Toolbox[4] in MATLAB R14, which uses the Virtual Reality Modeling Language (VRML) for scene modeling and display. This display will be available on the Base Station, enabling command personnel to monitor locations within a building. This will be of significant importance when hazardous situations occur. Information from sensors will also be displayed by the same software, using appropriate formats – e.g., for temperature (by color), sounds (by type, noise, voice), radiation or biohazards (by color), etc.

Map building: The second mapping algorithm takes precedence in the absence of floor plans. While first responder positions are collected and plotted, a 3-dimensional floor plan is automatically generated with a lifelike look and feel – as shown in Figure 9. This figure is a reconstruction (in real-time) of the path in Figure 1 using JAVA3D objects to build walls and corridors.

[pic]

Figure 10 Building internal structure constructed by JAVA3D map-building software

for the path shown earlier using data provided by the INU.

This capability is important in constructing hypothetical models of the interior of a structure. As individuals make their way through a building, the software attempts to identify key features such as stairwells, elevators, and hallways, and generates walls, floors and other features, to define the structure of the building to aid the incident commander viewing the progress.

4 Miniaturization of Prototype

Under the funding from MFRI, we have worked on developing second generation hardware for the Firefighter Sentinel system. The final design boards have been sent out and we will have prototypes for testing by mid-late April.

The motivations for development of a generation two system are:

• miniaturization

• improved heat dissipation

• addition of sensors

• improved compass

• replacing INU to data radio wiring with Bluetooth

• replacing analog sensors with digital counterparts

• connector upgrades

As shown in Figure 1, the Fire Sentinel Beacon in its new design has 2 pieces, the INU is worn on the firefighters belt. The INU is now connected to the data radio via Bluetooth, while the generation 1 required a wired connection. With the addition of the Bluetooth wireless connection, the radio can be connected in any convenient location so as not to interfere with movement or other gear.

The improved INU, which we plan to test at MFRI in May or June, has several new motion detecting microelectronic sensors. These sensors are made possible with the advent of MEMS sensors that have only become available in the last several months. The new INU includes a MEMS three-axis accelerometer, two MEMS gyroscopes, and a 3-axis magnetic field sensor. The data from the accelerometer will be used to identify individual steps taken by the subject, or other distinctive features of the motion. The gyroscope provides the rate of angular change of the user’s direction. This information will be used to help calculate the direction of the user’s line of progress. Data from the magnetic field sensors (compass) will also be used to calculate the direction of the subject’s trajectory.

These sensors are controlled by a dsPIC signal processing micro-controller chip. The data obtained from the sensors, will be combined with unique algorithms to ascertain the precise trajectory of a subject in motion. Currently, the algorithm is optimized for walking or stair climbing in buildings, and largely relies on the gyroscope to track variations in direction. With the new hardware, the algorithm can be enhanced to more reliably determine whether the subject is crawling, running, etc.

The role of the magnetic sensor data will be increased to provide another source of directional data in addition to that obtained from the gyroscope. Errors in compass readings resulting from nearby ferrous objects will be corrected by comparing the total local magnetic field magnitude to that of the earth to recognize corrupted compass data. The algorithm will be coded using assembly language into the dsPIC. This microprocessor, which is optimized for signal processing, will implement the algorithm, and output XYZ coordinates of the subject. Error correction algorithms will be developed to accommodate for sensor readings which are momentarily corrupted.

The dsPIC will calculate the coordinates of the user, and send them via blue tooth to the motherboard microprocessor. This information will then be incorporated into the node report that the motherboard assembles, and output to the transceiver. The transceiver will then transmit the node report, which includes the coordinates of the subject to other team members.

In summary, the modifications we are making to the system are:

• INU: We are miniaturizing the INU and adding several new sensors so that movements such as crawling can be tracked. We are adding a Bluetooth module for communications between the INU and data radio.

• Data Radio: we are combining the motherboard, GPS unit, and data radio onto one board and adding a Bluetooth module for communications between this unit and the INU.

1 INU Improvements

Figure 12 and Figure 13 show block diagrams for the first and second generation INU, respectively. Key improvements are the addition of z component to the magnetic field magnitude and the second gyroscope. These are critical components for tracking firefighters when they are crawling, or any other non-vertical mode of locomotion. The first generation INU has limited capability to track a user when they are not vertical. For this reason generation two has includes a vertically mounted MEMS gyroscope sensor and vertical axis magnetic field sensor. These additional sensors are necessary for determining the direction the user is crawling. The vertical axis magnetic field sensor will also assist in compensating for compass bearing errors.

|[pic] |[pic] |

|Figure 11. First Generation INU Block Diagram |Figure 12. Second Generation INU Block Diagram |

Analog sensors used for the INU have all been replaced with their digital counterparts. The new digital sensors operate using SPI (serial peripheral interface) which allows them to all connect to the microcontroller via a common three wire bus. This allows PCB trace routing space to be saved. In addition the digital sensors do not require buffers between their output signals and the microcontroller. By eliminating these buffers, a large amount of additional board space is saved.

The first generation INU is connected to the node controller via a 8-wire interface and was considered to be a standard plug and play sensor. For generation two, the wired interface was replaced with a Bluetooth wireless connection that is dedicated to communications between INU and node controller. This benefits the user as it is less awkward, the risk of snagging wires is eliminated, and time is saved because there is no additional connection for the user to make while putting on the unit.

For the second generation INU all component sizes were reduced. The package type for all resistors and capacitors was changed from 1206 (0.12” x 0.06”) to 0402 (0.04” x 0.02”). The accelerometer was changed from a 15.4mm x 7.5mm leaded SOIC package to a 7mm x 7mm leadless QFN package. The microcontroller was changed from a leaded 12mm x 12mm TQFP package to and 8mm x 8mm leadless QFN package. The gyro was changed from a 26.5mm x 11.3mm evaluation module to a leadless 8.2mm x 8.2mm LGA. The programming and power connectors were also replaced with smaller counterparts. Total PCB dimensions were reduced dramatically. The critical dimension that defines how far the unit protrudes from the body was reduced from 64mm to 14.3mm.

• Old INU dimensions 13mm x 81mm x 64mm

• New INU dimensions 37.6mm x 45mm x 14.3mm

|[pic] |[pic] |

Figure 13. Second Generation INU Boards a) INU main board is worn vertically with back against firefighter’s belt. b) INU secondary board inserts into the main board at silk screened rectangle and allows for tracking of firefighters in non-vertical orientations such as crawling.

Libraries have been created for new parts with manufacturing ease and heat management in mind. This involved defining solder masks openings and optimized solder paste stencil patterns for each component. In addition, thermal vias were placed beneath the microcontroller. These vias connect the exposed paddle of the QFN package to the ground plane and a copper area beneath the IC to dissipate heat.

Assembly instruction packages have been created for the INU which include formatted component centroid data file, silk screen and copper gerber files, solder paste gerbers, formatted bill of materials, and special instructions. This information is used by the board assembler to place the parts.

PCB design was panelized and a border was added to meet requirements for automatic component placement machines. Score lines were drawn and cut so that assembled arrays of PCBs can be easily snapped apart. Each board array contains 9 PCB’s. This not only eases component placement but reduces the cost for each PCB.

Mounting holes were placed on the PCB to allow for easy and secure installation in a plastic enclosure. Prototype enclosure will be made from flame retardant ABS plastic with flame rating UL94-0. Outer dimensions of enclosure will be about 2.5” x 2” x 0.8.

5 Node Controller

The first generation node controller played the role of reading data from external sensors (INU, GPS, etc.), formatting data packets for transmission to the base station and sending the packets to the data radio for transmission. The data radio and GPS were separate modules that communicated with the node controller via a 8-wire interface. This modular approach was taken in the beginning stages of the project so that different devices could be evaluated on the same system. We were able to determine from testing data radios and GPS which were most reliable.

[pic]

Figure 14. Second Generation Node Controller

Generation two combines the selected data radio and GPS module with the node controller on the same PCB. The new node controller will still have plug and play ports but the connectors to these ports have been improved to provide extra reliability. The second generation node controller has additional features including 4Mbit onboard memory for data logging, onboard temperature sensor, onboard audible tone generator for alerts, low battery detection with LED indicators, USB interface, alert button to signal base station for help, and dedicated GPS controller.

The microcontroller for generation two has also been upgraded to a signal processing microcontroller. This will allow for more efficient computations within the device.

As mentioned above, the first generation node controller was connected to the INU via a 8-wire interface, where the INU was considered to be a standard plug and play sensor. For generation two, the wired interface between the two devices has been replaced with a Bluetooth wireless connection that is dedicated to communications between the node controller and INU.

The new node controller will operate on four AA alkaline (or any chemistry over 1.3V/cell) batteries which will fit conveniently into PCB mounted plastic holders. All electronics on the node controller PCB except for data radio will operate at 3.3V. The data radio will operate at 5V. Additional mounting holes on the PCB will allow for easy integration into an electronics enclosure.

6 Future Work

Further development and testing of the Sentinel System is necessary in order to transition this technology into a product that is available for firefighters across the country. TRx Systems, Inc. and partner company Techno-Sciences, Inc., along with MFRI, have applied for further government funding from a variety of sources in order to continue this development. Elements of the future work are outlined below.

Signal Processing

Remaining key technical challenges are self initialization, (auto) calibration, and diagnostics. In addition, algorithms that automatically recognize and adapt to varying motion types and users must be developed.

Specifically, we shall develop “multi-mode” tracking algorithms that can “recognize” the nature of the movement (walking, running, crawling, etc.) and select the appropriate signal processing algorithm to compute location data for the context.

Mesh Networking

The mobile nodes in our sentinel system form a mesh network. By forming a mesh network data is allowed to be routed (“hopped”) though other nodes if the nodes cannot form a direct link with the base station. The ability to hop extends the range of the system. Our current system supports a single hop operation, which means if the data cannot get back to the base station directly; it can be routed through one other node to get there. We are in the process of upgrading the networking software to allow multiple hops, thus further extending the range of the network. In our testing so far in up to 4 story buildings, a single hop has been sufficient to obtain full coverage.

The mesh network also allows data to be communicated directly between mobile nodes on the network, so that a node can know where all the other nodes in the network are. Currently the nodes do not store information from other nodes. The redesigned mother board has onboard memory that will enable local storage of information on everyone in the network, so for example, each node will store the last position of all other active nodes.

Sensor Information

The conditions a firefighter endures in a fire fight require extreme physical fitness. But in the adrenalin rush of the fire, even a firefighter in peak condition can push his body too far. Stress, heat, high body temperature and dehydration are all factors that can conspire to cause critical injury. In fact, heart attack remains the leading cause of death among firefighters (44 percent of firefighter in-the-line-of-duty deaths are attributed to heart attack). A system that monitors and transmits a firefighter’s biometric data such as temperature, heart rate and blood pressure, could detect a firefighter in distress and provide necessary and timely information for their extraction.

Our motherboard microprocessor has the capability to receive data sent from blue tooth devices. We will demonstrate the capability to communicate with off the shelf heart rate monitors that send blue tooth signals.

The firefighters gear has become so good that they are not always able to feel the external temperature to know when they are in danger of overheating. Our INU hardware contains a temperature sensor and in addition we will add a sensor into the firefighter’s jacket. Using data from these sensors and an algorithm developed and tested by MFRI in conjunction with the University of Maryland Fire Protection Engineering Department, we will give the firefighters (and incident commander) a time remaining warning when the firefighter enters a high heat environment. The algorithm computes exposure levels and estimates tolerable heat levels based on a model of the insulating properties of the coat. As the temperature decreases, for example, from controlling the fire, the time remaining will increase.

Information to Firefighter

Our current system collects information (position, sensor data, …) from the firefighters and transmits it to the base station for display. We have not yet developed a system that allows the firefighter to get information from the command station, for example, for alerts to evacuate, or for egress direction, or for directions on most quickly get to a downed comrade. We will develop and test a varying systems for getting this information to the firefighter. The systems we intend to test range from simple systems including LEDS in the SCBA mask to indicate warning or direction, to preprogrammed text messaging and/or audio messaging (similar to the Magellan system in cars) to a blue tooth ear piece, to a fully functional HUD that would eventually be built into the mask. (See Figure 13 for a commercially available system, that is expensive .) Even a simple communications system will allow the incident commander to get important information quickly to the person or group that needs it.

[pic]

Figure 15 Liteye HUD mounted on a helmet. The viewer is “transparent” enabling the user to see objects ahead.

Getting user feedback on these communications systems is critical. We have seen devices with nice displays that we have been told will not be able to be seen in a smoke filled room, or with audio that we are told could not possibly be understood in the noise of a ranging fire. Testing our systems with MFRI and local fire departments will be a valuable step in coming up a workable design.

We expect that this system will be an add-on module to our main tracking device

Testing

Further testing and customer feedback are crucial. To accomplish this we will continue to work with the Maryland Fire and Rescue Institute (MFRI). With the help of MFRI our technology will be tested successively more difficult operational scenarios. First the system will go through trials in the burn building and smoke maze with fire fighter trainees in full turnout gear. Once these tests are completed successfully and feedback is aggregated and addressed, MFRI will arrange for local fire departments to beta test the system and provide feedback. In parallel, we will test the system in buildings that are complex from the point of view of transmission of radio signals. Test results and user feedback from will be a major driver for rapid improvement of system. Testing the system in a variety of different scenarios will help to assess and eliminate failure modes. Feedback from users will be rapidly turned around to improve system usability by the customers.

Certification

We have begun to look into certification testing. At this point there are no NFPA standards for location and tracking technology but our plan is to take what we can from the NFPA 1982 Standard on Personal Alert Safety Systems. We have spoken with Steve Sanders at The Safety Equipment Institute (SEI) who has made us aware of a developing electronics standard NFPA 1800 that may also have relevant tests but that will probably not be complete for another couple of years. In the meantime, we will follow the development to assure we are prepared for any new testing. SEI is a local company that provides third-party certification programs to test and certify a broad range of safety equipment products. The Safety Equipment Institute headquartered in McLean, Virgina. SEI's certification programs are accredited by the American National Standards Institute (ANSI) in accordance with the standard, ISO Guide 65, General Requirements for Bodies Operating Product Certification Systems.  We will have the system certified for Intrinsic Safety (UL 913). We plan to sub this testing to Intertec Testing Services in Courtland NY. This is also one of the labs used by SEI in NFPA 1982 certification testing. We plan to initiate certification design reviews before about 6 months into this effort to ensure that there are no clear issues with our electronics and enclosures that need to be addressed before testing can begin.

Manufacturing

We have identified and met with a subcontractor that we will be using for enclosure design and manufacture, EG&G in Ijamsville, MD. They have engineering as well as full CNC and conventional machine shop facilities to support our effort. In addition, they have experience with developing for and going through certification testing. We have identified companies to make and assemble the circuit boards. Advanced Circuits ( ), who we have worked with in the past, will make the boards and Screaming Circuits () will assemble them.  Both companies will do small runs and are also capable of running productions quantities. We plan to do final assembly and system testing in house for quality control purposes.

-----------------------

[1] . See also for floor plans.

A DVD with a larger format vers[2] ion of the video is available.

[3] The command station will have the ability to import maps and building floor plans as available at the incident.

[4] We are using MATLAB to develop the software design and key functions. The commercial version of the software will be compiled code that does not depend on MATLAB.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download