Project CG-29: iHear2, A Multi-Channel Microphone Probing System
Project Client: Deng Audio Research
Project Description: iHear aims to improve the quality of life for millions and provide hearing health accessible to everyone. By 2050, nearly 2.5 billion people are projected to have some degree of hearing loss, and at least 700 million will require hearing rehabilitation. Early detection is key to mitigating permanent hearing damage.
A recent invention by Deng Audio Research offers a more precise and efficient way to measure the acoustic parameters of ears. This development has significant applications in diagnosing auditory disorders, improving hearing device design, and enhancing headphone safety.
This detection is done by placing an acoustic tube inside the patient’s ear and incidenting sounds of different frequencies. The reflected wave is then recorded by a microphone array on the metal tube to calculate the acoustic parameters of the patient’s ears.
The innovative algorithm uses the reflection sound wave in a metal tube connected to the ear canal to calculate useful ear parameters with efficiency and precision. iHear2 employs a multi-channel MEMS microphone array inserted into the metal tube to detect sound wave reflection from the ear canal. The device can achieve high sampling frequency, high SNR, and a wide frequency range to ensure accurate and convenient recording for audiologists.
Project JM-58: Developing a Wearing ExG Device to Study Sleep
Project Client: SimPL Lab
Project Description: As group JM58 of the 2023-2024 capstone year we’ve designed and developed a full fledged application that monitors sleep for UBC’s SimPL Lab. The goal of the project was to design and implement an application to visualize, record and analyze EEG, ECG and EMG data from multiple amplifiers in one easy to use format, such that researchers can quickly and efficiently study sleep instability in subjects with a greater level of standardization.The major challenges of this project were analyzing and displaying real time filtered data from multiple third party amplifiers, performing complex post processing functions to this data and creating an application to house these various functions in one easy to use package.
Project JM-67: Backend Development for Motor Deficits Detection Using Hand Drawing and Facial Expression
Project Client: NeuroPrior AI
Project Description: Our project aims to develop an online website platform that will offer a scalable, secure, and efficient solution for analyzing data from hand motion and facial expression tests from users that suspect or are under ongoing motor deficit treatment. By providing a prediction percentage, we help users estimate how likely they are affected by motor deficits diseases.
This backend system of the website provides a robust platform for real-time data analysis, secure data storage, and seamless integration with the existing Human-Computer Interaction systems, i.e., website services.
This backend serves as the backbone for our client’s ongoing and future research, allowing for an in-depth analysis of human behavior and interactions. Moreover, the project assesses the scalability and compatibility of integrating more future services in the field of online user data analysis, thereby aiding our client in staying ahead of the technological curve.
Project JM-82: Smart Glove for Osteoarthritis (OA) and Rheumatoid Arthritis (RA) Populations
Project Client: TruMotion Technologies Ltd.
Project Description: Through working alongside TruMotion Technologies Ltd., capstone team JM82 has designed a smart glove to combat the symptoms of osteoarthritis (OA) and rheumatoid arthritis (RA) in the hands. RA is an autoimmune disorder targeting synovial joints, causing inflammation and damage, while OA results from gradual cartilage degradation due to mechanical stress and aging. These debilitating conditions cause pain and difficulty of movement in the hands, and affect the daily lives of over 32 million people in the United States alone. Treatment of the associated symptoms commonly include compression or heat therapy which help alleviate joint stiffness and discomfort, however, TruMotion aims to provide an all-in-one solution to obtain the benefits of both treatment options.
The smart glove utilizes specially made silver-coated nylon actuators to provide the mechanism for combined heating and compression, which are wound around afflicted areas of the hand to contract and stabilize the joints while simultaneously delivering heat. Through intermittent actuation, the glove provides compression of up to 25 mmHg of pressure and temperature delivery of up to 40°C, with an estimated battery life of 5.5 to 11 hours depending on the desired strength of the treatment. The Smart Glove’s output is further monitored and regulated through the use of pressure and temperature sensors, allowing for a high level of control and safety measures.
Project JY-68: Development of a Wearable Sensor Device for Human-Robot Interaction with EEG Capabilities
Project Client: NeuroPrior AI
Project Description: EEG (electroencephalography) devices unobtrusively monitor the electrical activity of neurons in the brain. These devices detect deviations in brain activity and may indicate the presence of a neurological disorder, such as insomnia or epilepsy. Traditional EEG sensors often come with high price tags and intricate operation mechanisms, limiting the scope of people with access to EEG tests.
Our team has designed a wearable EEG device which serves two main purposes:
- Our client is currently developing an AI algorithm which detects erroneous EEG signals. Our product will allow our clients to easily and efficiently collect data to train this algorithm.
- The wearable EEG device will allow medical practitioners to utilize remote appointments, where the patient can record brain signals at their own home and upload the recordings to be analyzed and will allow for the doctor/neurologist to form a diagnosis.
Our wearable EEG device is designed to be compact, affordable, and user friendly. We fit a host of EEG hardware into a battery-powered wireless package with a footprint no larger than a toonie. EEG data is sent over a Bluetooth connection to a computer, where waveforms are displayed on an intuitive website accessible by patients and medical professionals alike.
Project JY-69: CalorieCam HM- Development of a Smartphone App to Estimate Human Milk Calorie Content and Optimize Nutrition in Preterm Infants
Project Client: UBC/BCCHRI
Project Description: CalorieCam-HM is a mobile application that analyzes the colour of human milk to estimate the calorie content of the milk. By knowing roughly how many calories are in a sample of milk, nurses are able to supplement the milk with the appropriate amount of nutrition fortifiers to feed to preterm babies and ensure that the babies are getting the nutrition that they need.
Visually comparing a milk sample to existing colours in a palette can be quite subjective due to the slight differences in human perception. Our application solves this technical challenge by standardizing the colour comparison using image processing algorithms. These algorithms extract the colour values from an image taken by the phone camera based on the current lighting conditions to ensure that the comparison is consistent.
Project JY-92: Wearable System for Health Status Monitoring
Project Client: UBC Adaptive Microsystems Lab
Project Description: Overcrowding in the hospital emergency department (ED) has been a significant challenge within the Canadian healthcare system. The ED at Vancouver General Hospital (VGH) suffers from staff shortages relative to the high volume of visiting patients, leading to extended wait times up to 8 hours and reduced patient care quality.
Our wearable patient’s health monitoring system is capable of monitoring heart rate, blood oxygenation, breathing rate, and temperature, tracking body positions such as standing, sitting and lying, detecting seizures and falls as well as recording patient general location. The primary objective of our low-cost system is to provide medical personnel with a tool for continuous, real-time observation of vital signs and patient behavior. This will enable the medical staff to allocate more time to assess and care for patients in urgent need while also protecting patients. Our wearable health monitoring system is able to alert staff of unexpected deterioration in patient status due to abnormal vital signs, falls, seizures and no movement for a long time which can significantly increase the chances of receiving timely medical care in unexpected critical situations where every second matters for survival.
Project SF-36: A Lab in a Shoe: Detecting Trips and Stumbles with a Wearable Device
Project Client: University of British Columbia
Project Description: Walking serves as our main form of mobility, but as we age, our ability to coordinate and avoid obstacles decrease. This leads to an increase in the chances of stumbling or even falling when walking. To better understand these events in the normal environments of everyday walking, we worked to develop a solution that was able to measure the clearance of a step at various points of the walking cycle. By leveraging a combination of time of flight sensors, gyroscopes and accelerometers, we aimed to create a system that could accurately assess the spatial relationship between the walker’s foot and the ground. These sensors provide precise distance measurements, while gyroscopes and accelerometers help track the orientation and motion of the foot throughout each step. Together, these components enable us to capture detailed data on step clearance, allowing us to identify potential issues with gait and balance. Our goal is to provide this information for the UBC Sensorimotor Physiology Laboratory to further develop interventions and assistive devices that can help elderly maintain their mobility and reduce the risk of falls.
Project SF-41: Human Anatomy Teaching App
Project Client: UBC Faculty of Medicine
Project Description: Dr. Majid Doroudi, an Associate Professor of Teaching in the Faculty of Medicine at UBC, is committed to helping students master their knowledge and understanding of human anatomy. To this end he has invested time and resources into producing high-quality educational videos that explain anatomical concepts to students around the world. He has also taken the initiative to create and maintain the UBC Human Anatomy Teaching App, which is available on iOS and Android devices. This app is a pedagogical tool that was created out of the need to offer students a tool to learn anatomy visually through videos, quizzes, and flashcards that contain professional photos and diagrams of the human anatomy. The purpose of our project was to revive and improve this app.
The UBC Human Anatomy Teaching App hadn’t been updated in over 2 years. As a result, the app became incompatible with the current Android operating system and had broken features that no longer worked properly on Apple devices. It also lacked certain features that Dr. Doroudi had wanted to add to the app, such as the ability to delete uploaded content and add audio narration to flashcards and quiz questions. A third problem Dr. Doroudi had been encountering involved the app’s backend server, which had been periodically going down, requiring him to contact the developers of the app to reboot it. This was not sustainable for Dr. Doroudi and led him to discontinue use of the app in his teaching of anatomy courses.
Our team’s contribution to addressing the design challenge involved redesigning the user interface of the mobile application to support the current release of Android and lineup of iPhones while ensuring flexibility to new design changes in the future. It also included updating the admin portal application to offer Dr. Doroudi, or any app administrator, more control over the content being stored, as well as a modernized user interface. A last challenge we solved was to increase the reliability of the server such that it does not require frequent reboots and is easy for Dr. Doroudi to reboot by himself when needed.
Project TL-32: Use of Frequency Modulated Continuous Wavelength Radar (FMCW) to Detect Heart Rate, Respiratory Rate of Patients in Hospital Waiting Rooms, Seclusion Rooms and Those in Police Custody
Project Client: Aberrant Designs Inc.
Project Description: Emergency departments see significant morbidity and mortality each year from unrecognized changes in patient vital signs. A patient’s condition can deteriorate after they are first admitted; this is indicated by changes in heart and respiratory rate, which are difficult to notice visually.
Patients in hospital seclusion rooms are especially at risk, since traditional wired health monitors cannot be used. Despite having camera and in-person examinations to assess the patients, accurately distinguishing between a sleeping patient lying in bed and one experiencing distress – and quickly intervening to prevent life-threatening or long-term harm – remains a critical challenge.
Our solution uses a Frequency Modulated Continuous Wave (FMCW) Radar to monitor the heart and respiratory rate of a patient in a seclusion room in a non-invasive manner. FMCW Radar allows us to measure the movement of a patient’s chest with sub-millimeter precision, without the need for wires and contact sensors.
Our product processes the radar data in real-time; this includes signal processing algorithms to denoise and reconstruct patient vital signals, making our system more tolerant to patient movement and applicable to realistic conditions. We send the measured heart and respiratory rate values to a web-based GUI, allowing health workers or nurses to monitor multiple patient vital rates remotely. Additionally, the GUI creates audio and visual alerts when patient vitals are in a dangerous range, improving patient outcomes.
Project TL-43: Gesture Control and Vital Function Detection Using 60GHz Radar Sensing and AI/ML Processing
Project Client: Delta Controls Inc.
Project Description: Obtaining occupant metrics within a room, such as the presence or number of people within it, whether anyone has fallen, or if a medical crisis is occurring, was before only able to be done – autonomously – by using a camera. This introduces privacy concerns, especially in public settings. Using non-intrusive, highly accurate radar sensors provides an alternative. Delta Controls and our capstone team have utilized time-of-flight (ToF) and frequency modulated continuous wave (FMCW) sensors to facilitate touchless gesture control of room environment, track occupant numbers, detect occupant falls, and monitor heart and respiratory rate. This is done without the use of a camera or any intrusive optical imaging. Our team leverages machine learning and signal processing algorithms to decode radar data into useful occupant information that can then be used by a building management system (BMS) to streamline room control, trigger alarms in event of emergencies, and reduce dependency on human staff – all while respecting the privacy of occupants.