STEREO CAMERA OBJECT DETECTION FOR AUTONOMOUS VEHICLES
Motor-vehicle collisions result in thousands of deaths each year. Although not all collisions can be prevented, many can be avoided using self-driving vehicles. Autonomous vehicles require the use of computer vision to detect objects. This project uses stereo cameras to detect and classify objects for an autonomous vehicle. The main advantage of stereo camera based object detection, instead of relying on the driver’s vision, is that an autonomous system is not susceptible to fatigue, sickness, intoxication, or distraction like a human.
Team members: Matthew Magni, Mike Sililiane, Gourave Verma, Alton Lau
NIJI: WEARABLE HEALTH MONITORING AND TRAINING DEVICE FOR PETS
Niji is an affordable pet-wearable device that assists owners in monitoring their pet’s health and establishing training routines. A smart harness measures biometric data and communicates with a companion app on the owner’s smartphone. Owners can monitor their pet’s health data and receive notifications on the pet’s daily routines. Intelligent algorithms are used to detect abnormalities and alert the owner as necessary. Niji enables owners to establish routines for their younger pets, while simultaneously monitoring their health concerns as they grow older.
Team members: Ryan Chu, Alan Qu, Yi Sui, Lloyd Torres, Quan Zhang
DEAFBAND
DeafBand is a headband-like device that can recognize and pinpoint the direction of certain sounds (e.g., car horn, emergency alarm). The device is an integrated system of microphones and vibration motors that process audio input and notifies which motor on the band to trigger. Certain settings can be altered via a Bluetooth-paired smartphone device. DeafBand’s benefit over existing alternatives is its ability to directionally pinpoint the source of the sound, whereas other products can only output proximity.
Team members: ChangWoo Park, Raymond Zhuang, (Dan) Dongoug Kim, HanBat Kil
CONDUIT: AD-HOC COMMUNICATION NETWORK FOR SMARTPHONES
Smartphones are ubiquitous for communication but quickly become useless when out of cellular range. This lack of communication can be dangerous, especially in remote areas. Walkie-talkies remedy this situation, but it is not possible to achieve the level of functionality we are accustomed to with smartphones. Conduit is a system that uses alternative communication methods to serve as an ad-hoc channel for smartphone data. This is achieved by using radio-frequency transceivers interfaced to smartphones to transmit rich data, including text messages and GPS coordinates.
Team members: Aaron Lobo, Shadmaan Mahfuz, Alvin Joys, Davan Basran, Navjot Panesar
VIRTUAL EYES
Virtual Eyes is an obstacle-detection system for visually-impaired people. The system is embedded into a walking cane and incorporates various sensors to provide information about obstacles around the user. The system provides haptic feedback through vibrators on the user’s fingertip positions. As a complement to a traditional walking cane, it provides additional information about surroundings without affecting the user’s valuable senses such as hearing. The solution is more portable, is affordable, and provides more intuitive feedback compared to existing products.
Team members: Yuhao Deng, Zi Li, Cheng Guo, Songlin Yang, Yuanping Xia
NOSTRAIN: A CARPAL TUNNEL SYNDROME PROTECTION WEARABLE
In Canada, 15% of the population suffers from Repetitive Strain Injury (RSI), the most prevalent RSI of the hand being Carpal Tunnel Syndrome (CTS). This condition is common in people involved in repetitive wrist motion. The objective of this project is to design a lightweight wearable to detect common posture-related root causes of CTS by tracking hand motion. The device sends user data from the sensors to a mobile application that analyses this data and provides corrective measures to prevent further strain on the wrist.
Team members: Tharaga Balachandran, Smruthi Murali, Naima Khan, Shashank Sabhlok
WATERLYTICS
Water is a vital household need and essential for survival. However, around 50% of household water supplies fail at least one aspect of the Guidelines for Canadian Drinking Water Quality. The objective of this project is to design a water-analytics platform for consumer homes that monitors water-quality data such as pH and conductivity levels. In addition, the platform collects water-consumption data to provide insight into consumption patterns and harvests energy from the water flow to power electronic components.
Team members: Zuhaer Zim, Pela Cefalas, Shail Shah, Jawad Amin, Jian Feng Gao
SLEEPYZZZ – AN INFANT SLEEP TRACKER
SleepyZzz is a monitoring system that aims to track an infant’s temperature, movement, and heart rate while they sleep. The system is wirelessly connected to a parent’s mobile application, which can show the baby’s vitals in real time. Should the infant develop a fever, roll over on their stomach, or have their heart rate drop, SleepyZzz will notify the parents via their mobile device so that they can react and mitigate the problem as quickly as possible.
Team members: Donald Ngai, Adam Seeto, Maxwell Carter, Ioana Babus, Andy Kwok
SIXTHSENSE
SixthSense is a wireless sensor network geared towards consumer use in applications such as home automation, biomedical sensing, and robotics. Each sensor node is either an inertial measurement, light, infrared, or temperature sensor that is easily mounted to many surfaces with an adhesive backing. Sensor data is wirelessly transmitted to a central hub where it is aggregated and then displayed through a web application. SixthSense allows consumers to easily build their own wireless sensor networks by eliminating the complicated setup required by existing industrial solutions.
Team members: Daniel Beaudreau, Harsh Shah, Laurence D’Alessandro, Vidyasagar Rajendran, Stefan Jovanovic
POWERCYCLE
Due to traffic congestion and air pollution in metropolitan cities, there is a need for cheap and eco-friendly transportation. The project objective is to design an energy storage and conversion system which works within a public bike share. Each bicycle carries a device that charges a battery during journeys. Upon completion of a journey, the bicycle is docked at a station and the battery’s stored energy is transferred to the grid. Reward points, which users earn for completed journeys, are used to offset rental costs.
Team members: Stanislaus Kelvin Hidajat, M.Fahad Faizan, Syed Hammad Hasan, Asef Mustakim, Suchen Wang
PIXUP: IMAGE ENHANCEMENT
Photography plays a significant role in people’s lives. It allows people to capture a moment and share it with the world. Many photos, however, suffer from noise and digital artifacts such as blurriness, aliasing, ringing, blocking, and color banding due to the low quality of cameras in many mobile devices. PixUp allows users to remove these artifacts from their images in a simple and user-friendly way.
Team members: Tom Peng, Zikang Zhao, Chengyi Zhou, Jiaqi Wu, Danheng Shen
50SENSE
According to a Stanford study, 90% of crashes are caused by human error. The objective of this project is to fuse data from two sensors (an IMU and a GPS) to establish a vehicle’s state relative to a reference frame. The project implements an Extended Kalman Filter using the Ackerman Steering model to perform sensor fusion. The main advantage of this solution over manual driving is that several robust sensors work in unison to establish the vehicle’s states, whereas a human is prone to distraction and unpredictable variability.
Team members: Karanbir Randhawa, Malcolm Mathew, Anthony Luu, Azad Memon
EXCILIA: WEARABLE TO HELP IN SPEECH THERAPY PROCESS
Excilia is the world’s first speech-therapeutic wearable device for children to detect speech anomalies and provide real-time monitoring capabilities for both the parent and the speech therapist. The speech pathologist is able to proactively monitor the effectiveness of their speech therapy and have personalized therapy sessions based on real-time data gathered outside the session. The advantage of this design over alternatives is the automatic detection of speech anomalies for real-time feedback to therapist and the child.
Team members: Dhruvit Gandhi, Kartheek Gajjala, Mohammed Baquir, Niral Fernando
UNREAL SENSORSIM: SENSOR SIMULATION SUITE FOR AUTONOMOUS VEHICLES
Unreal SensorSim is a software-in-loop simulator, built on the Unreal Engine physics simulator, for autonomous vehicle systems. From the autonomous vehicle system, the vehicle control output move vehicle models in the game engine, and the graphical output is rendered on a cluster and relayed back to the vehicle software for autonomous driving calculations. The advantage of this design is the low latency throughput which allows the vehicle software to learn in real-time scenarios.
Team members: Yubo Han, Yichen Huang, Eric Lomore, Mike Zhou
AUTONOMOUS SERVICE CART SYSTEM
Automation of the food service industry is becoming a reality. However, the method of food delivery still typically relies on servers manually bringing food to individual tables. The objective of this project to design a robot to automate the delivery of food in à la carte restaurants. The robot uses multiple cameras for localization and travels to designated locations upon request. The robot also gathers information on the dishes ordered for analysis. This project is aimed to reduce service shortage in busy restaurants.
Team members: Yifan Wang, Bonnie Tseng, Frank Zhang, Bobby Li
I AM HEAR WEARABLES: HEARING IMPAIRED NOTIFICATION ASSISTANT
The objective of this project is to improve communication with those who are deaf or hard of hearing in a family or social setting. The system behaves like a pager system to alert the deaf or hard of hearing individual that their attention is needed. The device comprises of embedded electronics in the form of a processor, modem, accelerometer, vital signs sensor, and Bluetooth. Advantages of this design over major alternatives are ease of wearability, increased functionality, and simplicity of the interface.
Team members: Dylan Lau, Jenna Godwin, Peter Kim, Akram Aoun, Shao Hua Zhang
AUTONOMOUS VEHICLE MAPPING SYSTEM
The Autonomous Vehicle Mapping System is designed for mapping and routing autonomous vehicles. This system provides map viewing and location querying capabilities as well as route calculation including turn-by-turn communication with the vehicle. The advantage of this system is that the custom turn-by-turn directions produced can integrate seamlessly with other subsystems of the vehicle, where other mapping competitors do not support autonomous vehicle integration.
Team members: Uzair Haq, Alex Henderson, Ming Yan, David Lee
PARK-N-WALK
Park-N-Walk is an automatic parking payment system that securely and precisely charges drivers based on real-time parking information. It consists of tags installed in vehicles, street-side readers that detect tags when they are in a parking spot, and central controllers that update parking records in a database server. The main advantage of Park-N-Walk is that it removes the annoyances associated with carrying coins, understanding how to use parking meters, correctly estimating time, or rushing activities to adhere to a prepaid time limit. Just park and walk.
Team members: Jenny Yu, Serina Tan, John Xu, Evey Liu
AUTOMATED SPICE DISPENSER
When cooking, making accurate measurements for each spice is challenging and time consuming. Often multiple clean utensils are required when preparing spices. The Automated Spice Dispenser is created to eliminate these issues by automating the process of preparing spices without the need for any utensils. The device allows the user to input specific quantities of spices via a smartphone app or an onboard screen. It then dispenses all spices in the correct quantity simultaneously, saving time and any error in measurements.
Team members: Ayush Kumar, Hitoshi Kumarawadu, Rakesh Mandhan, Acer Wang
SIDEWALKER: THE PEDESTRIAN SAFETY DEVICE
Pedestrians are frequently startled or injured in collisions as they increasingly share sidewalks with cyclists and longboarders. Sidewalker’s objective is to design a pedestrian safety system that perceives approaching traffic behind the user and provides alerts for incoming danger. Using an integrated microwave and camera sensor system, acquired data is transmitted via Bluetooth to the user’s smartphone for signal and image processing. While collision avoidance systems on bikes rely on slow universal adoption, Sidewalker’s unique advantage is enabling pedestrians to improve their own safety.
Team members: Rafik Dalati, Tony Wu, John Adeyemo, John Xu, Austin Wong
HEADIT VR
Heading is a very effective way of scoring, delivering, and defending the ball in soccer. While an important skill to have, it can lead to concussion and cognitive degradation, especially when practiced in succession. In some youth leagues, heading is banned due to these risks, which are more serious for children. Using Virtual Reality, motion tracking, and vibration feedback, HeadIt VR provides an alternative to the conventional heading practice, and allows children and injured athletes to practise without the risk of injury.
Team members: Hang Lyu, Jenny Xiang, David Park, Ken Lin, Andrew An
CAMTRACK: YOUR PERSONAL AUTOMATIC CAMERA BOT
The objective of CamTrack is to design an autonomous, mobile camera quadcopter that can track the movement and direction of an object, to allow solo video producers to film without a dedicated camera operator. The drone tracks the target using image processing to determine the relative distance and direction of the user. The solution makes use of control theory and robot dynamics for smooth video tracking, and multivariable control systems for camera positioning.
Team members: Boyang Zhang, Hsin Je Ho, Jason Cheung, Alastair Yu, Kelvin Li
AUTOMATIC TRANSMISSION FOR A BICYCLE
With a greater number of cyclists every year, there is an increased need for bicycle products to enhance the cycling experience. This project aims to design an automatic gear shifter that can easily retrofit a standard bicycle. The device adjusts the gear ratios on the bicycle automatically, factoring in the cadence of the rider’s pedaling and the slope of the terrain. In any weather, it can be used to reduce both the physical impact on the rider’s joints and the wear on the bicycle’s chain.
Team members: Peter Pogorski, Amanda O’Meara, Conrad Montor, Sean Buller, Kris Wagner
VIRTUAL REALISKI
Have you ever wanted to ski while it’s 20 degrees outside? Virtual Realiski is an immersive skiing simulator that fulfills just that need. The system uses sensors to read a user’s body movements and posture, and replicates them in a virtual reality simulation. The platform also tilts in response to the slope’s gradient to give the full experience of skiing without worrying about frostbite.
Team members: Nikhil Ramburrun, Denny Hoang, Vivian Mac, Hamoudi El-Imam
POSTURE DETECTION SYSTEM
A common issue in office environments is poor sitting posture, which can lead to many chronic problems. The Posture Detection System monitors a user’s posture and notifies the user when they deviate from proper posture. It consists of a suite of force sensors and a camera. Currently, solutions to poor posture include ergonomic chairs and frequent physiotherapy sessions which can be costly. The aim of the project is to provide a convenient solution for the user to improve posture, at an affordable price.
Team members: William Healey, Xun Yin, Aniket Limaye, Harvey Xu, Armin Motekalem
MAGIC MIRROR
Magic Mirror offers online retail consumers a virtual fitting room with which they can try on clothing items without leaving their home. The Magic mirror combines commodity hardware with advanced software to create an improved experience for consumers in the digital age.
Team members: Sadman Khan, Kevin Yang, Ryan Perry-Nguyen, Christopher Ciufo, Zheng (Andy) Liang
SENSENET: SELF-DEPLOYING SENSOR ARRAY
SenseNet is a dynamic sensor network that can be used to autonomously monitor areas for sensor events. SenseNet uses robot agents connected to a centralized server that optimally controls the robots to respond to events and monitor targets. The advantage of this design over static sensor networks is that SenseNet can dynamically reconfigure itself as robots and targets are added or removed.
Team members: Sam Walters, Stephen Molyneaux, Stewart Walsh, Kieran Yee, David Murdoch
BIKEPIN: BIKE TRACKING AND THEFT DETECTION SYSTEM
BikePin is a bike theft detection and location tracking solution enabling bike owners to react proactively to their bike being stolen. It detects if there is motion when the bike should be stationary and pushes a notification to the owner that suspicious activity has been detected. The advantage of this design is that it provides bike owners a better method of detecting when their bike is being stolen and offers more information of where the bike may be.
Team members: Yuki Ji, David Vuong, Tiffany Lui, Matthew Di Nardo
EASILIGHT
The EasiLight LED system aims to incorporate a home’s lighting system with various smart-home functions into one customizable network that is completely accessible from a smartphone. The fixture connects to a home Wi-Fi network and allows the user to control lighting behaviour, stream audio, and collect data via the custom android application. Furthermore, EasiLight can also be used locally or remotely which allows for peace of mind when away and makes life easier when at home. EasiLight makes customizing, controlling, and monitoring one’s home easy.
Team members: Elvin Li Kam Wa, Abigail Miner, Howard Wei
WHO SAID IT
Reliable speech recognition in the presence of noise and multiple voices is a problem faced by existing speech-to-text solutions. Who Said It is a system that uses a blind source separation (BSS) algorithm to filter out one person’s voice from noise and other voices. Once the BSS algorithm has separated the voices, they are displayed on a screen for the user to read. This device allows hearing-impaired individuals to participate in a conversation in the presence of background noise and other voices.
Team members: Michael Carlson, Henrietta Odiete, Chien An Liu, James Whiteside
PILLMINDER
In the US, suboptimal patient prescription adherence rates cost $290 billion for the health care system annually. PillMinder helps to alleviate this issue by providing a reusable and affordable solution that prompts patients to take their prescribed medication. PillMinder accomplishes this by combining a mobile application with a WiFi-enabled container that can reliably track and remind patients to take their scheduled medication.
Team members: Yifei Deng, Sangyoun Kim, Wenxi Zhao, Allan Wang, Kevin Chen
GKPRISM: THE OBJECT CLASSIFICATION AND FACE DETECTION PATROLLER
GKPrism could be considered as the base of an autonomous patroller. It aims to perform real-time facial detection to find people of interest by periodically patrolling through a pre-defined route. The onboard computer fuses data from cameras and distance-detection sensors to produce the relative locality and classification of the surrounding objects. The face detection system can identify people’s faces and upload these faces with a timestamp and location information to the server, which can be used by the end users for future reference, e.g. for forensic identification.
Team members: Arthur Cheung, Desmond Lin, Kai Luo, Chris Zou, Wenjing Chen
LITWEAR: A VITALS MONITORING SYSTEM FOR FIREFIGHTERS
This project is designed to improve the safety and efficiency of a firefighting crew in a building-fire environment. The system consists of a wearable device used to monitor heart and body-heat activity. The wearable communicates wirelessly to a base station that performs digital signal processing on the raw sensor data. All the data is presented to a base station operator who can use the information to make informed decisions in real time.
Team members: Dmitrii Gluscencov, Kevin Wikant, Kyle de Kock, Emmanuel Utuama
FLEXEAT: DINE-IN MADE EASY
FlexEat is an application that allows a user to directly order food from their mobile device at participating dine-in restaurants. FlexEat also caters to users with dietary restrictions so that they can convey their needs directly to the chef and reduce the risk of any allergic reaction. Restaurants will benefit from FlexEat over traditional food ordering systems by reducing the need for waiters, thus reducing the restaurant’s staffing costs. FlexEat is a universal application which can be used at multiple restaurants.
Team members: Mandish Shah, Rushil Master, Andrew Chung, Chaitanya Khanna
DRONE EMERGENCY RECOVERY SYSTEM
Drones have become an increasingly popular tool for capturing aerial photography. However, drone crashes are a serious concern as they can lead to the loss of expensive equipment. The Drone Emergency Recovery System uses accelerometers, gyroscopes, and the drone’s flight data to detect when a crash is imminent, then automatically ejects a parachute and activates a locator beacon to ensure safe landing and easy recovery of the drone.
Team members: Wei Du, Naihe Feng, Jimmy Zhang, En Wei
GAZEIFY: ANDROID SDK TO HELP APPLICATION TRACK USERS’ EYE MOTION
Mobile app designers are highly empathetic, wanting to better understand their users to achieve better in-app experiences. This project equips designers with a mobile analytic solution that can deliver more useful information on how users are interacting with content. Gazeify is a software-service solution that uses the phone’s existing hardware and deep-learning models to track what users look at to give designers a clearer picture of how users interact with their designs. It’s as simple as plug-and-play.
Team members: Peter Pengda Sun, Hung Yui Lo, Ariel Paulin, Rongqiang Cai, Boran Wang
YOUCAI: GROCERY SHOPPING APP
There are over 36,400 restaurants in Ontario alone, generating about 3.9% of its GDP. Bulk food suppliers tend to focus on specific ingredient categories, and enforce a minimum amount per order. This prevents restaurants from maximizing their potential profits. The solution, an app written in React Native, helps consolidate all types of groceries, allowing restaurants to easily order the ingredients they need. Furthermore, the app provides analytics that allow users to gain insight on their spending and product trends in the market.
Team members: Maidi Lin, Malcom Li, Kai Jun Weng, Timothy Tong, David Yu
PIANOSCAPE: THE FLEXIBLE AND COST-EFFECTIVE PIANO TUTORING SYSTEM
Pianoscape is a piano tutoring system that provides real-time feedback to a user’s performance. It allows students to learn music sheets/scores of their choice by importing scores of different formats into its system. It uses Optical Music Recognition (OMR) for scanned scores to extract music elements from images. This project was motivated by the need for a less costly way of learning piano, and a more available system for continuous feedback. Due to Pianoscape’s flexibility, users can progress their learning with even standardized piano books.
Team members: Ekteshaf Chowdhury, Shayon Chowdhury, Sam Li, Matthew Ngai, Nick Ticknovich
GPU PARALLELIZED PATTERN MATCHING
Pattern matching is an essential technique for evidence retrieval in digital forensics. It aids investigators in searching through various data storage media for specific keywords relating to cases. However, these searches are time consuming, especially when performed on hard drives at sizes of 1 terabyte or more. The objective of this project is to speed up this process by designing an algorithm that takes advantage of the parallel computing capabilities of Graphics Processing Units.
Team members: Alex Chien, Zeyu Wang, Alex Tong, Haicheng Zhan, Bill Ding
ECKO: AUGMENTED REALITY AUDIO TRANSCRIBER
Ecko is a real-time audio transcriber application implemented on augmented reality glasses to aid the culturally deaf and hard of hearing. On the glasses, Ecko transforms speech from a wireless microphone into text which is displayed in a closed captioning fashion. Ecko is a better alternative to the traditional hearing aid as the wearer can visually see live transcribed text near the speaker instead of simply hearing amplified audio. This also offers a convenient method of understanding people who cannot sign their speech.
Team members: Yuchen Huang, Kia Shakiba, Tasnim Alam, Marieta Farova
FLUTTER: VIRTUAL PIANO WITH HAPTIC FEEDBACK
Flutter is a virtual piano system that recognizes hand gestures and provides contactless haptic feedback. Efficient dynamic gesture recognition is accomplished with the use of a series of optical sensors and machine-learning algorithms. At the same time, a sensation of touch is accomplished by focusing ultrasonic waves with an ultrasonic transducer phased-array system. The Flutter system, compared to other available virtual piano systems, offers low-latency gesture recognition and accurate haptic response, mimicking the feel and response of a real piano.
Team members: Martin Li, David Li, Lisa Zhang, Jimmy Hung
ADVANCED TABLE TENNIS AUTO-SERVER
The advanced nature of this automated table tennis server is its ability to perform a variety of serves and receive feedback about how the player responds to each serve. The design uses an air compressor to launch the ball and motors to serve the ball with an appropriate trajectory. Sensors continuously scan the table to capture when the user returns the ball. The advantage of this design over others is that it adapts to the user’s playing style, creating a more realistic experience.
Team members: Partheban Paranthaman, Asha Arudchelvan, Bo Wang, Ethan Greavette, John Feather
BIOACOUSTIC
BioAcoustic is a computer-aided auscultation device that helps healthcare professionals with the detection of common heart and lung ailments. The device uses sensors around the chest area to capture and amplify sounds, converts them to digital signals, and analyzes the signals to detect possible conditions. The advantage of BioAcoustic over current alternatives is that it reduces human interpretation while also being less invasive and expensive than advanced medical equipment.
Team members: Faraz Mahagavi, Abdullah Navid, Avininder Maan, Ali Ajmine, Steven Ming
COLLABR
It is important to allow software developers to be able to collaborate with each other. The project objective is to create a text editor to allow developers to collaborate with one another on code remotely. The system provides a simple way for developers to collaborate with their peers, allowing text documents or program files to be edited in real-time. In comparison to current collaborative editing tools, this project is a standalone desktop program while other implementations are web-based applications or plugins.
Team members: Steve Davis, Daniel Chow, Albert Le, Jason Salaber
WATTMAT
In Ontario, approximately 60 people, many who are older adults, visit the hospital every day for injuries resulting from a fall involving ice and snow. WattMat is a smart heating mat system that automatically melts ice and snow build-up when detected. The system analyzes criteria such as temperature, energy cost, weather forecast data, and accumulation of ice and snow to achieve maximum efficiency. The advantages WattMat has over existing alternatives are its ability to optimally supply heat, its low cost, and its overall power efficiency.
Team members: Roushd Ahmed, George Abraham, Hassaan Naeem, Vatsal Patel, Derek Trider
NURSE JOY
Nurses are overworked, causing hospital patients to experience delays in receiving help. Nurse Joy is a wheeled robot that delivers items such as food, drinks, and medical supplies to bedridden patients. Starting from a dispatching station on the hospital floor, Nurse Joy navigates to the designated patient’s bedside and delivers the item requested. This reduces the workload for nurses and improves the quality of service for the patients.
Team members: Roger Zhang, Xingchao Yin, GunWung Lee, Alan Zhang, Leo Kim
ACCULUS
The Acculus is a computer-interaction system that accurately and affordably tracks a stylus’ position and rotation in 3D space while providing positional and rotational haptic feedback. The system works using a robotic arm with a set of encoders for exact positional tracking and a gyroscope for precise rotational tracking. A developer API is provided for integrating the Acculus into any application easily, as well as providing historical data for activity analysis. The Acculus makes realistic virtual interaction possible.
Team members: Zach Waterfield, Samuel Simpson, Alexio Del Gobbo, Jackson Prange
LHITEBOARD
Whiteboards are a great way of sharing ideas. Electronic adaptations such as interactive touch surfaces build on whiteboard functionality with additional features, but have portability constraints, are fixed in size, and are prohibitively expensive. Lhiteboard enables any flat surface to be used as an interactive surface using a projector and infrared (IR) tracking, all at a low cost. Input is provided with an IR pen and, through image processing, the system computer responds accordingly, whether drawing on a virtual canvas or navigating an operating system.
Team members: Joseph Truong, Wa-Sang Tse, Eduard Permyakov, Hyunjun Yoo
CONVOYING FOR SHIPPING COST REDUCTION
Surface freight transportation is predicted to grow dramatically. Presently, 60% of all land shipping is done by Heavy-Duty Vehicles (HDVs). When convenient, groups of HDVs can travel closely in a convoy, minimizing their aerodynamic drag and reducing fuel consumption up to 25%. The objective of this project is to design an algorithm that calculates an optimal path, maximizing convoying to achieve greater fuel efficiency than simply taking the shortest path.
Team members: Sean Amadio, Hazheer Foroutan, Geoffrey Heath, Aditi Lath, Priya Bibra
THE COPPER CHEF
Many variables affect the quality of a cooked steak, which should finish within a narrow range of acceptable internal temperatures, and have a good, even sear on each side. To the amateur chef, the perfect steak is rare and elusive. The objective of this project is to design and build a device capable of automatically cooking a steak to the user’s desired doneness, tracking the internal and surface temperatures of the steak, flipping it when necessary, and removing it from the heat when cooking has finished.
Team members: Taylor Laekeman, Brian Truong, Shiva Ehtezazi, Eric Bender, Christian Petri
PRISM
In the modern music industry, live performances are the main source of income for artists. However, smaller bands have a more difficult time providing a similar concert experience as that of more well-known artists. Prism allows up-and-coming artists to easily coordinate more visually immersive performances on a smaller budget. To accomplish this, Prism features a software application and wearable device that allows musicians to coordinate lighting scenes and automate spotlights while performing. As such, artists are able to design larger-scale performances with a smaller budget.
Team members: Filip Vranes, Rahil Hirani, Tyler Adams, Luka Pavlovic
CHORUS: INEXPENSIVE VOCAL RECORDING SYSTEM
Recording high-quality audio can be prohibitively expensive for amateur singers, given the high costs of professional studio equipment. However, inexpensive microphones, such as those in cell phones and laptops, are widely available. Chorus records audio using one or more inexpensive microphones and processes it using source separation and a convolutional neural network to produce high-quality audio. The users can coordinate recording and download the resulting tracks through a web interface. This provides a cheap and accessible recording solution.
Team members: Muntashir Hossain, Avi Menon, Trevor van Leeuwen, Alex Labach
WATSIGN
WATSign is a real-time classification program which provides real-time feedback based on the traffic signs present within the environment. Input camera feed collected from multiple sensors and cameras capturing the surrounding road conditions are fed into the edge detection subsystem which is then processed through the subsystem and passed into a neural network classifier subsystem. The advantage of this design is that WATSign provides, at an affordable price, real-time classification to mitigate accidental injuries and fatalities caused by automobile accidents every year.
Team members: Tej Nagabhatla, Amrit Parmar, Keshav Kanatala, Neesarg Patel
SERVACE
In tennis, return of serve is one of the most crucial components of the game and also the most difficult to practice. The ServAce is designed to mimic a variety of tennis serves, both in speed and spin, in order to help players practice their return game. It also has the ability to deliver unanticipated serve types to the user, replicating the uncertainties of a real match. ServAce is remotely controlled by an app so players can practice anywhere, even without a partner.
Malcolm Maloney, Hoyman Chen, Paul Soladoye, Abdullah Monoar, MinJun Park
TICKT: THE GROUP MOVIE RECOMMENDATION AI
Tickt is a movie recommendation app that helps groups easily discover new movies based on shared tastes and preferences in real time. Through considering everyone’s interests and watch history, Tickt intelligently recommends a movie that the whole group will love. It remembers the movies that a user has watched, and continuously learns, adapting to the user’s tastes. Tickt’s advantage is that it learns how to recommend movies for a group of people, rather than just for a single individual.
Team members: Ramie Raufdeen, Victor Szeto, Eddie Zhang, Yong Eun La
EYEDRONE
EyeDrone is a quadcopter system that takes in voice commands from a user to detect if a certain object of interest is present in the field of view from where the drone is situated. EyeDrone then indicates to the user if the object is present or not. The design of this project required the use of image recognition, controls, networks, and natural language processing.
Team members: Fahad Syed, Hamza Azeem, Daniyal Alam, Huzaifa Shafiq
BILLIARD BUDDY
Billiard Buddy is a tool that provides users with optimal shot paths to practice their skills, keeps track of progress, and allows them to become the best players they can be. Using image processing to translate existing game states to a virtual table, players interact with different balls allowing a proprietary algorithm to provide multiple paths to sink the chosen ball. Billiard Buddy’s advantage is being able to modify any billiards table and providing paths to sink the ball whereas other tables only provide trajectories.
Team members: Umesh Senthilnathan, Alex Ohn, Samin Salam, Shahriar Syed, Johnathan Ho
THE WIRELESS CHARGING TABLE
The Wireless Charging Table charges a phone through a table, regardless of the phone’s location on the surface. The project consists of a detection system, which involves a grid of infrared sensors to pinpoint the phone’s location. A motor system moves the wireless charging device to a location under the phone by means of controllable motors and a rail system under the table. Existing wireless charging solutions require the user to manually place the phone on a charging device.
Team members: Jimmy Nguyen, Omar Doo, Sina Abedi, Akkas Mughal, Daniel Richardson
FORTEPIANO: THE PIANO SHEET MUSIC TRANSCRIPTION SYSTEM
FortePiano is a system that transcribes a live piano performance into sheet music with the corresponding finger mappings and dynamic annotations used, so that another musician can learn performance techniques more efficiently. It uses audio processing and image processing to transcribe the notes and the musical nuances during a performance. The advantage of this software-centric design is that FortePiano facilitates ease of use during transcription and annotation, allowing the pianist to play naturally without wearing any additional hardware devices that are obtrusive to playing.
Team members: Wei Xin, Joanna Chu, Kevin Truong, Yuxuan Chen, Bei Cong Zheng
PROJECT WATNEY
Maintaining a garden is a daunting task for the inexperienced gardener. The objective of this project is to design a system that monitors the environmental conditions of a garden and informs the user of plant needs. Project Watney accomplishes this with a network of sensors nodes that measure the state of the garden and transmit the data to a central controller. The results are combined with information about each plant's optimal growth requirements and presented to the user through a web application.
Team members: Alex Williams, Milind Shah, Gleb Billig, Guozhen La
AUTOMATED ACCESS-GRANTING SYSTEM
The proliferation of internet-connected devices around us has created plenty of opportunities to make our homes smarter. Home security, in particular, could benefit from advances in machine learning and image recognition. Specifically, the objective of this project is to design a system to independently identify visitors to a house, be it friends or family, and grant them access primarily using their faces. This system brings the power of deep learning and face-recognition technology right to your doorstep.
Team members: Sanchit Gera, Varun Batta, Joshua Leo-Irabor, Irvanjit Gill
RAKSHA: STREET CHILDREN TRACKER
In the developing world, there are countless non-profit organizations that are trying to help over 100 million street children out of poverty. However, they are unable to effectively allocate their resources without insightful data about these children. Raksha is a wearable device that periodically measures and transmits a child’s location and body temperature, and is equipped with a panic button to request for help. The data is available to social workers to provide insights into where to efficiently allocate resources.
Team members: Yash Kothari, Jeffery Yaw Owusu-Ansah, Janujan Selvaratnam, Eric Hollbach
HOLOGUIDE
Approximately 500,000 people were injured while exercising in 2012 in the U.S. The objective of this project is to design a cost-effective educational system to prevent injuries caused by improper form during exercise. The system combines data from depth, thermal, and infrared sensors along with cameras to capture an expert performing exercises with the correct technique. The 3D model is rendered and displayed next to gym equipment in the form of a holographic display to demonstrate correct technique.
Team members: Steven Boseovski, Sashakt Tyagi, Aman Kumar, Milan Dimic, Dishant Patel
EXOMUSCLE
There is a lack of technology to actively protect from musculoskeletal injuries. Competition in athletics causes individuals to demand more from their bodies, increasing the likelihood of injury. ExoMuscle provides smart sensing joint sleeves for enhanced athletic performance and injury prevention. By integrating sensors and actuators into compression fabric sleeves, one’s muscular exertion is tracked and utilized to provide active feedback and support. The system is able to provide the optimal amount of support to maximize performance improvement while minimizing risks of injury.
Team members: Casey Garcia, David Xu, Jonathan Manson-Hennig, Runze Gao, Odelia Fung