Project Name: Development of a wearable device for pregnant women to track fetal activity
Project Client: Flutter Wear Inc.
Project Description: Flutter Wear Inc. is a Vancouver-based organization with the goal of developing wearable technology for pregnant women, in order to accurately track and record fetal activity. Their core mission is to improve the maternal experience by allowing mothers to feel connected to their developing fetus, thereby easing anxiety and facilitating improved mental well-being. Typically, a mother manually keeps track of fetal movements to ensure the baby is developing healthily. However, it may be difficult to distinguish between fetal activity, bodily convulsions, and other movements in the region. Furthermore, fetal movements can exert varying force, so a mother cannot consistently feel all movements, consequently leading to higher anxiety and concern. To address these issues, Flutter Wear aims to create a comfortable and stretchable device that can be worn on the mother’s belly. Our capstone team developed a device which uses robust ECG and force sensors in combination with enhanced signal processing algorithms to detect fetal activity, and transmit them to a mobile app in real time. The app will track these movements over long periods, thus allowing mothers and practitioners to analyze the patterns and determine the health of the baby.
Project Name: IoT Smart Home Presence Detection
Project Client: iaconicDesign Inc. (Modena Smart Home)
Project Description: Our client, iaconicDesign Inc., is dedicated to designing smart home devices that have sleek and elegant aesthetics for their Modena Smart Home line of products. Conventional presence detection devices rely on light-based technologies which require the use of intrusive lenses to improve range and accuracy. For this project, our team evaluated and prototyped alternative presence detection solutions that are small and unobtrusive so that they can be used in the Modena Smart Lightswitch. Based on our research, our team decided to use micro-electro-mechanical systems (MEMS) microphones: sensors that can generate electrical signals in response to vibrations from sound propagation. Our team has developed an acoustic human presence detection algorithm that uses data from a MEMS microphone that can be hidden inside the lightswitch enclosure, preserving the stylish design. The algorithm is designed to detect voice activity and impulses, such as footsteps. While the algorithm is designed to be highly configurable, we have done extensive testing to decide parameters that minimize false positive and false negative results from outside noise.
Project Name: Risk Model to Determine the Possibility of Injury of People Utilizing an Engineered Playing Field Adjacent to a BC Hydro 138 kV Overhead Transmission Asset
Project Client: BC Hydro
Technology: High Voltage Electrical
Project Description: BC Hydro has power lines all across BC responsible for over 95% of electricity delivered throughout the province. Ideally, there should not be gatherings near these lines for extended periods of time. Our project explores a playing field located in close proximity to a power line support structure. Due to the increasing number of gatherings at the playing field each year, there was a need to assess the possible risk of injury from electric shock to the users of the field. Our team is tasked for the risk assessment of the field. Using VBA, we developed a tool in Excel to help our client quantify risk due to electric shock for various types of public gatherings. The tool allows for a wide range of inputs and models how factors such as soil conditions and number of people present on the field relate to the risk of electric shock. Our solution will help our client implement appropriate safety measures for a variety of scenarios.
Project Name: Feasibility Study to Determine the Best Design Alternative to Obtain GPR Step and Touch Potential Data during Typical High Power Short Circuit Qualification Testing of Transmission and Distribution Equipment at a Test Facility
Project Description: This project is a feasibility study for Powertech Labs, a high voltage testing lab, that will determine how viable it is to integrate heat and pressure sensors into their testing facility. Our objective is to develop a data acquisition method that collects arc flash data from everyday short circuit testing, specifically for large air gaps, which in turn will lead to better identification of arc flash safety zones. This study is significant to our client because with improved identification of arc flash safety zones, Powertech will be able to help BC Hydro, the main electricity distributor in the province, in assessing the hazards to their workers at substations and in the power line environment.
Project Name: Feasibility Study to Determine the Best Design Alternative to Obtain Arc Flash Related Data during Typical High Power Short Circuit Qualification Testing of Transmission and Distribution Equipment at a Test Facility
Project Description: Our client, Powertech Labs, the R&D subsidiary of BC Hydro, is concerned for the safety of their personnel as they are constantly exposed to dangerously high levels of electricity. At the lab, operators run many equipment-failure tests per day. In the event of a failure, the surrounding ground becomes energized. This is dangerous for nearby workers as this can cause electricity to flow through their bodies. Powertech Labs wants to first identify the level of risk that occurs while running their daily tests before performing any mitigation actions. Our team is assigned the task to introduce a real-time detection and measuring system that can capture and record these electrical hazards that occur throughout the facility. With this measuring system, our client can ensure the safety of their workers. Technical challenges involved in this project include modelling the client’s lab facility, simulating a failure test, and identifying the high-risk locations based on electricity levels and high traffic areas. From the study, specifications for the detection system (i.e. measurement range, sampling frequency, etc.) are identified. The proposed solution is designed in accordance with the Canadian Electrical Code and is to be interconnected to the client’s existing data acquisition system.
Project Name: Developing a wearable head impact IMU to study sports concussions
Project Client: SimPL
Technology: Computer & Electronics
Project Description: Sub-concussive head impacts are becoming a more pressing health issue year by year. Different from concussions, these are impacts that rarely display conclusive symptoms but can potentially lead to long-term neurodegeneration as a result of successive impacts accumulating over time. Collision events during sports is one such way a person can get a sub-concussive head impact injury. The Sensing in Biomechanical Processes Lab aims to develop a wearable technology that can be used to measure and capture the sub-concussive impacts and to conduct further studies on their fundamental biomedical processes.
With the eventual goal of being housed inside a sports mouthguard, our hardware design is the third and latest iteration of the series of prototypes. By placing the components onto two separate printed circuit boards, a great deal of flexibility is achieved. This also allows for a design with a lower overall complexity, cost, and easier accessibility when it comes to maintenance and troubleshooting. The device uses the onboard accelrometers to keep track of the forces that an athlete experiences throughout the course of a game. If the athlete experiences a significant head collision, the device would be able to detect this through a spike in accelerometer readings and save it as a datalog in the onchip flash storage. When connected with the accompanying mobile application, the datalog can also be downloaded wirelessly via Bluetooth and analysed by the researchers.
Project Name: An Apple Watch app for clinical trial follow-up in congenital heart disease patients
Project Client: Vancouver Stroke Program
Project Description: Congenital heart disease (CHD) consists of various types of anatomic heart problems present from birth. It generally increases the risk of stroke in both adults and children. The study of adult congenital heart disease (ACHD) is becoming increasingly prevalent as the number of children with CHD surviving into adulthood is rising. The Vancouver Stroke Program is piloting a study to evaluate the risk of cerebrovascular disease (CBD) in the ACHD population. They are trying to establish a correlation between strokes and cognitive decline in the ACHD population. By determining the correlation between strokes and cognitive decline, the study aims to serve as a basis for identifying patients who may be at a greater risk for CBD and who may benefit from dedicated neurological follow-up and targeted preventative strategies. We have provided the VSP with a tool in the form of an Apple Watch and an iOS app, which will improve participants’ long-term engagement with the study. This application will allow the research team to collect a larger set of data by streamlining and automating data collection and will also allow for additional information that would not have been collected through a simple phone call or online form.
Project Name: Software for detection of cancer cells in pathology images using smart phones
Project Client: BC Ovarian Cancer Research Program (OVCARE)/Vancouver Coastal Health Research Institute
Project Description: Cancer Cell Detection aims to solve the inequality in access to specialized cancer healthcare for Canadians living in remote communities. Specialized healthcare is centred around major urban centres in Canada. Currently, members of remote communities must travel to these urban centres to see an oncologist, or have a biopsy taken at their local clinic and mailed to a specialist for diagnosis and treatment. Our system is designed to modernize this process. A microscope technician at their local clinic captures an entire biopsy slide with our mobile app and a microscope. Our stitching algorithm combines these images into one complete image of the entire slide. A machine learning model will then perform the specialized task of detecting tumorous regions within the slide. The technician then captures these tumour areas at a higher zoom level in more detail and sends all photos of the biopsy slide to a specialist instantaneously. By leveraging modern software technology, our design will cut down on the time and expenses incurred by Canadians during a stressful time in their lives.
Project Description: The Boatswain Boat Tracking and Monitoring solution aims to address the lack of sufficient monitoring solutions for vessels at sea or in harbour. Alongside our client Sierra Wireless, a global leader in cellular IoT technology, we explored the capabilities of Sierra’s data orchestration cloud platform, Octave, and their open-source IoT board, the mangOH Yellow. The Boatswain hardware is housed in a splash-proof enclosure and includes a custom-designed PCB that retrieves data from a suite of on-board sensors. The mangOH Yellow’s cellular connectivity allows the boat’s sensor data, such as the boat’s GPS location, bilge activity, and shore power connection, to be communicated to Octave. The user can access all their boat’s critical assets from the easy-to-use Boatswain mobile app. Users can also elect to receive immediate push notifications in the case of unexpected events. These events include the boat moving out of a user-defined location zone, the bilge pump running for more than a specified number of minutes, the boat getting accidentally disconnected from shore power, or an intruder entering the main cabin. With Boatswain, recreational boat owners, marina owners, and boat rental service providers can easily monitor their vessels with a compact and comprehensive solution.
Project Name: Machine Learning and data fusion approaches to assess sleep EEG in neuro-degenerative disorders
Project Client: Pacific Parkinson’s Research Centre, Djavad Mowafaghian Centre for Brain Health, University of British Columbia
Project Description: Sleep stage analysis is a key factor in the diagnosis and disease severity prediction of multiple neurodegenerative disorders. This analysis is typically performed using EEG signals (brain waves) recorded in-hospital via polysomnography (PSG). Recently however, there has been a shift to collecting EEG data at home with specialized headbands. This method is much less expensive and is more comfortable for patients, but the data collected by these headbands is lower quality than that produced via PSG – causing automated sleep staging algorithms to perform poorly. Our capstone project, put forward by the Pacific Parkinson’s Research Centre, is to help level this trade-off by creating an automatic sleep staging solution which performs well with headband EEG data. We have successfully developed a deep learning system which obtains a high accuracy on both low-quality two-channel EEG headband data and with the gold-standard PSG data. We have also authored a research paper detailing our methods and findings in our search for this solution.