Department of Computer Science and Engineering ...



1. Any project listed here: . Android Tablet “Button Builder” and Communicator, to aid speech therapists and their disabled clientsCommunicative disorders are varied and age-independent. Many people, children included, lack or lose the ability to speak, yet maintain fine motor control. An Android tablet app that displays buttons for phrase selection could be a portable voice, customizable for different situations and locations. The device app must have two modes, one for speech therapists to build button pages, and one for the client to use as a communications device.Build: buttons must be added to a page, and pages must be added to the app. Pages merely contain button layouts. Buttons should be customizable as to size, location, color, action, page (including multiple pages), label, and picture. Button Actions should include speaking a word or phrase, or moving to a different page. The app itself should be customizable to male or female voice, and volume. Use: Pages should be presented, and pages and buttons should be easily selectable. When a button is pressed, the preprogrammed action should be taken. Suggestions: In addition to speaking a word or phrase, buttons might also be capable of playing a song or video, or controlling an electrical outlet to activate a device.3. Audio Situational AwarenessThe Desktop Audio Situational Awareness device monitors the surrounding environment for active audio sources including speech and other audio events. The device is intended to enable the Deaf and Hard of Hearing (DHH) to know when someone is speaking, what they are saying and where they are located in group discussions such as in the classroom, workplace and at social gatherings.Audio Situational AwarenessAudio Situational Awareness provides spatial detection to identify and associate successive audio signals from the same source to present a consistent view of the audio environment. Audio sources are detected by a microphone array and the source direction is determined using Time Delay of Arrival (TDOA). The difference in TDOA of audio signals detected at different microphones in the array is used to determine the direction of the audio source relative to the microphone positions.Microphone array beamforming is used to capture the audio signals originating from a detected source. Audio signal events from speech sources are translated to text while other audio sources may be identified by their distinct characteristics.Desktop Audio Situational Awareness DeviceThe Desktop Audio Situational Awareness Device is similar in design to an Amazon Alexa device. The device case will be 3-5 inches in diameter and ? inches thick. Thinner and smaller is better for portability. Rounded edges are desired on both bottom and top with vent holes around the perimeter sufficient in size on the bottom for cooling (if necessary) and on the top for audio detection. There are two controls for power on/off and audio detection on/off (mute).24314158509000The device electronics comprise a single All-in-One motherboard (MB) with an internal battery as a standalone power source. A cost/ production design tradeoff may be to use a general purpose MB (e.g. Raspberry Pi) with a separate daughter board for the microphone array. This may impact the diameter and thickness of the device.The All-in-One MB includes a low-power CPU (e.g. Arm Cortex A7 1.5 GHz quad-core processor with 1 GB memory), an array of six (6) MEMS microphones equally spaced around the perimeter, Bluetooth Low Energy / Wi-Fi (802.11n), a micro SD slot, and a USB port for programming, external power and charging. Wireless (Qi) charging is desirable. The Respeaker Core V.2 is an example of a third party Alexa style MB.The battery power supply should be sufficient for at least 4-6 hours of continuous operation, although a longer period of 10-12 hours is desirable. Wireless and fast recharging capabilities are also desired.System RequirementsThe ASA will use the ReSpeaker Core vers. 2.0 as its hardware platform.C++ will be the programming language of choice for the ASA.Listen ModeThe ASA shall monitor the surrounding audio environment for speech sources. This will be called listen mode.The ASA shall be able to detect speech audio. The goal is to detect speech up to a range of 30 feet in a classroom environment.The ASA shall determine the location (direction of arrival) of a speech source relative to the device.The ASA shall track active speech sources by location.The ASA will track active speech sources as they move relative to the device.A speaker identifier shall be assigned to each speech source.Multiple speech sources collocated within 10 degrees of each other shall be considered one speaker. If there is a way to distinguish between collocated speakers, they may be assigned individual identifiers.The ASA will convert speech to text for each speech source. A wireless connection to the Internet is permitted to achieve this. The ASA shall start processing speech when the device is powered up.User InterfaceThe user will address the system through a Bluetooth connected device, preferrably an Android tablet or phone.The ASA shall transmit the speaker ID, location and text results to a Bluetooth connected device. Strong consideration should be given to encrypting this over-the-air transmission. The tablet shall log the transmission record in a timestamped text file, displayable on request. To enable debug capability, all Bluetooth-routed messages shall also be available on a monitor, and all Bluetooth-sourced commands shall also be sourced from a keyboard, connected directly to the ReSpeaker Core. Record ModeThe ASA shall record an active speech source for a fixed period. This will be called record mode.If a fixed period of silence (no active speech) is detected, the ASA shall stop recording and return to listen mode. The recorded speech audio shall be translated to text using an online Speech-To-Text (STT) service.Configuration ModeThe user shall have the ability to control the ASA device via the Bluetooth connected device.The user shall have the ability to save a speech transcript to a file.The user shall have the ability to reset or clear a speech transcript.The user shall have the ability to select the language to be translated.The user shall have the ability to select the language to be displayed.4. Buffalo and Erie County Botanical Garden Tribute Database Statement of Need BackgroundThe Buffalo and Erie County Botanical Gardens has a wide selection of dedications that remember love ones, honor special people, offer gratitude and/or commemorate special occasions. Gifts have been given to the Gardens to celebrate marriages, births, retirements, anniversaries, or memorialize a beloved friend or family member. In return for the gifts, the Gardens creates permanent dedications in the form of engraved bricks, pavers, stone monument, a plant, bench etc. These dedications are located both inside and outside on the grounds of the Botanical Gardens.In order to track and manage the Tribute collection, The Buffalo and Erie County Botanical Gardens compiled pictures of current Tributes at the Garden and a Google Doc file that provides information on each Tribute Type (e.g., brick or paver), a Tribute reference #, the Last name on the Tribute, the First Name on the Tribute, the Full Tribute Text, the Tribute Location, the Date of Purchase, the Purchaser Last Name, the Purchaser First Name, the name of the Photo of the Tribute and the Photo Title.Statement of Need (in order of importance)The Buffalo and Erie County Botanical Gardens would like a means (e.g., a software package or database with User Interface) to be developed that allows authorized staff members to easily update and maintain the data. This includes adding new tributes and associated tribute information including pictures, and to update tribute data if corrections are required or if tribute location is changed. Expandability of the data to include new fields in the future is also desired. In addition to staff who require editing access, the front desk staff requires viewing access in order to provide accurate Tribute information to visitors.The Gardens would like to also like to provide viewing access to the Tribute data (eventually to be accessed through our web site) in a mobile friendly fashion. This access would allow viewers to intuitively browse the complete selection of tributes and/or to search for specific tributes of interest. Ideally, completed searches will show a pop-up picture of the tribute and selected data (e.g., text on Tribute) but not purchaser’s name or any other confidential information. A map option is desired to allow staff and external viewers to see the approximate location of the Tribute so they can easily find any tributes of interest.The Buffalo and Erie County Botanical Gardens would appreciate the opportunity to incrementally review any Requirement Specification and Top Level User Interface Specifications that are conceived to allow staff input prior to the capability being developed.5. The Firefighters / First-Responder Locator and Monitoring SystemThe #1 cause of firefighter death during an emergency call is heart attack. There is a clear need to enhance the safety of our First Responders by continuously monitoring their position and sensing their vital signs and environment while in dangerous situations. A wireless networking device attached to each person could do this, as well as serve as an alternate communication channel to his or her voice radio. The device would provide them with three major advantages:Location: Incident Commanders, back-up personnel and the Responders themselves will always know their precise positionMonitoring of vital signs and environment: pulse rate, body temp, air temp, CO2 gas, etc. are sensed and communicated to nearby monitoring stationsBack-up two-way communications: panic button and other important information can be communicated as an alternate, robust path to the standard voice radio.This system must provide the following:The ability to monitor, track, and archive each firefighter’s location and vital signs (heart rate, oxygen level, temperature). Location information must include the entire path each firefighter took to achieve his/her current position.A 3D real time visualization of the incident scene, including the location of floors, walls, hot-spots, firefighters, and victims.The ability to click on and retrieve vital signs info for each firefighter on the scene6. A Talking Head for elderly or home-bound patientsThis system would provide a computerized attendant for people who live alone – simply an on-screen intelligent face with personality to greet patients every morning, prompt them for activity and monitoring (such as taking a diabetic blood glucose reading) and then react appropriately to the results (such as calling a doctor, or video-conferencing in the client’s family).When done as a prototype, people were very reluctant to give their assistant up, they had integrated it into their daily lives. Since Google and Amazon both make their home assistant APIs available, this can integrate Alexa, etc., although that is not necessary. A visual presence is necessary, coupled with a doctor’s ability to write speaking scripts and gauge answers.A touch screen with prompts and selections is allowable (“Please select your blood pressure range today”). 7. Intelligent Cosmetic / Fashion / Makeup Mirror There are many different apps that can help one see what they could look like if they change their look but there are none that will help instruct them how to change their look in real time with pin point accuracy of what they’re trying to go for. Everyday, somewhere, someplace, someone is always trying to perfect their personal image. People find themselves going to a cosmetologist or a hairdresser for their woes. But what if there was a better way? A way for people to save money and perfect their image theirself. Perfect application and dressing every single time. An app that will give you the ability to change your look before you decide to in real time and then instruct you how to create that look perfectly everything single time. No more mess, no more fret, no more wasted money. We have the opportunity with this tool to shake the foundation of cosmetics that will revolutionize the industry which has remained analog for a long time while helping the consumer have a new sense of confidence and self esteem they have always wanted in the privacy and safety of your home. See yourself as the perfect you all the time. This software is taking it a step up from regular makeup apps that only show what they could look like. We introduce the step of helping a person gain that look they want. This can first scan the user’s face while creating a 3d recreation of their face to help the software find every area so as to better help apply cosmetics. While keeping an image of what the user wants to look like, the software can red circle areas that require the user’s input. The darker the red, the more makeup needs to be applied, the lighter, the less. This can also help the user in weather they should move away from a center area or towards. As the cosmetic is applied well in a general area, the lines will fade away. Outside of this, the software can help best detect the color the user wants compared to what they have and can either recommend a specific color product they don’t have or how to best optimize what they do have.Outside of just facial cosmetics, this can help user’s with their hair in either cosmetics of that or styling it and instructing them how to in real time. Especially more so considering the current environment.And drilling down further would be an option for people to see how they would look with clothes before buying them. The app can take the user’s measurements with those of the item maker’s product and in real time or pictured show the user what clothes would look like on them before ordering anything, thus saving both the user and company they’re buying from save money. With this, the user is given the ability to always have the best look that they want in their eyes first and foremost. Considering most people are audio/visual learners, this will help them considerably at perfecting their look and style, all the while saving them money.This can be used in a multitude of ways. For starters, it can be a standalone app on either just any ios/android device though preferably a tablet. For extra computing power, it can be setup with a server cluster to help a mobile device do the computations needed to perform in real time. And finally, a proprietary device can be made as a professional “magic mirror”Coupled with hardware behind the mirror such as cameras, various sensors, mics, audio, and lighting, this gives the user more uses for the item beyond basic data such as time and weather.The idea is to craft a quality smart mirror that can be built at a low enough cost and still be a quality product. Taking into account that the basic magic mirror uses a raspberry pi wouldn't be realistic due to price, size, and features. A custom pcb built for use of only needed accessories with most likely an arm processor and at least 4gb of ddr4/5 ram to mass produce would be smaller and cheaper to mass produce. Possibility for a battery within but it could be hard wired if need be for it’s electrical needs. The biggest piece of hardware will be the one way mirror that can also be used as a touch screen and have a non smudge surface. Under that mirror will be a screen, dynamic lighting, camaras, mic/speakers, and various sensors for the purpose of the software being able to analyze the user for cosmetic help in real time. The light itself under the mirror should be a thin ring around the screen under the mirror that turns on when the user likes it and from there the user has options to brighten, warm, or cool the light. The mirror should also have speakers and a mirror for any use that the user needs. It could be hearing music or following along with a video or app helping them in this process and a mic for hands free use with an assisted ai program.The camera itself is still to be determined to make sure it works best with the machine. At least 12 megapixels, if not bigger. In regards to privacy with the camera, it will only be on when the user wants it and alert them to it being on. Besides that, for an extra layer of security helping the user feel more comfortable, there will be a small analoge tab at the top of the mirror that if the user wants to use the camera has to lift up because while down, it will be blocking the camera.For general power connection, the device shall have a usb type c cable with an adapter but besides that, we're looking at if the device should have a battery internally but that could always be relegated as an extra feature. And finally the device should have bluetooth capabilities so as to be able to connect other devices to it that can benefit the user experience.For this proprietary device, if constructed, it would need the following:1 1080p monitor1 8gb Raspberry PI1 Stereo camera2 normal cameras2 ir sensors1 Raspberry pi extender1 usb type c dongle1 micro hdmi to hdmi1 ir touch frame over lay1 two way mirror8. Mechanic access websiteOne thing in today’s day in age is finding a proper mechanic. People have a hard time finding a mechanic that can fit their budget, time, is honest, or all of the above. For this, most people are subjected to either trial and error or word of mouth. What if we could change all that. What if we could bring customers and mechanics together. I’m proposing an app that can and will allow customers to be able to find and talk with mechanics while also helping keep these transactions honest. Customers will be able to see reviews and what others think of a shop. In turn, shops will want to build their business and also their reputation.Think of this like a cross between lending tree, yelp, and bumble. Customers can put in their information and their problem while seeing shops based by them with things they offer and specialties they have. A customer can partially drill down what kind of work they need and help narrow the search. From there, customers can ask whatever their heart desires and mechanics can get back to them instantly(lots of shops have a person running a front desk). This app could be used on IOS/Android while also being web based. A user puts down their area and other user information, from there they can either see a map or listings of mechanics near them. They can drill down those searches by what the shop specializes in (ie forign cars, electric cars) and what benefits shops have (rental car, shuttle, ext). Another interesting part is having the shops show their prices for work, items, and other things they may have. The user would also be able to open up a query to communicate with the shop to ask questions. If a shop is picked, another interesting part is that they can be kept up to date with the shop on their cars through text, picture, and video.This would be all available through an ios/android app and a web based site for both user’s and shops. To help keep them honest and keep the system from abuse, shops can only be rated if a job is accepted and completed. With this we can also do a payment system like instacart where the transaction goes through us. Any job accepted, the app takes a percentage of.User usage - For the user, they would either download this app to the mobile platform of their choice or use the site. The user would have to create a profile to be able to use the services offered. Once downloaded, the user can ask to their heart’s content whatever they want and be responded to in real time from shops that the user can decide upon. If they pick a shop, from there, they can view the progress on their car in real time as well.Shop usage - Shops and especially smaller shops can now see a wider array of potential business. They can also now be more connected to see what customers want. This allows smaller shops to better serve customers and know what they want in order to compete with larger ones. They can help customers with every step of the way and also be a second opinion for customers as well.Testing - For this it would be creating the app and taking it to mechanics around UB while having students test it. ................
................

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

Google Online Preview   Download