AutoMate: A physiological fatigue detection system for drowsy driving prevention

Sleeping behind the wheel is one of the leading causes of road accidents and there currently exists a lack of definite limits on fatigue. The United States of America’s National Highway Traffic Safety Administration cites that over 100,000 road accidents in 2017 can be directly attributed to drowsy driving , and when over 1 in 2 Ontarian drivers admit to driving drowsy in the last month, it’s easy to see why this can be so dangerous. In fact, going 24 hours without sleep is equal to having a blood alcohol content of 0.10 %. That’s 0.02% above the legal driving limit.

Robust, Efficient, and Scalable Control of Hybrid Energy Systems usingArtificial Intelligence Planning

In the energy industry, as a result of global warming, population growth, and environmental, political, and
economic considerations, a fundamental shift in technology is expected. To address this, in this project we
propose to test the applicability of our Artificial Intelligence technology for solving a challenging computational
problem in the energy sector. We anticipate that our approach can offer significant benefits over currently
employed techniques.

A Window into the Future: Dye?Sensitized Solar Cells in Buildings

Solar cells are an incredible source of renewable energy. As with all technologies, though, they should be cheap,
easy to adopt, and smart. Dye-Sensitized Solar cells are solar cells that are not only highly effective at creating
energy from the sun but are also capable of being adopted into homes in new and unique ways. We are developing
a method of converting existing windows into solar cells. Throug this technology a window would still act as a
window, except it would also be capable of harvesting energy from the sun and powering a home.


A novel design of football helmet to mitigate the concussion is proposed here. The jerk transferred to brain, due to the collisions between players and due to falling on the field, will be reduced, attenuated, decomposed, and directed away from reaching the brain. Multi-shell made of composite materials, along with speed-dependent padding between head and inner shell will be used. Relative motion between shells will be attained though a specially designed structure made of strong material bars. Therefore, the force of collisions will be controlled by a number of safety layers.

Monitoring System for “Flushable” Consumer Products in Urban Wastewater Collection Systems

This research project explores the application of an artificial intelligence-based monitoring system comprised of image-based sensors and processing algorithms to detect, identify, and monitor the incoming presence of wet wipes and nonwovens in urban drainage systems in near real-time to pre-empt the effects of the damages caused by users’ disposal of these products in toilets. The AI-based system, to be employed in a number of monitoring locations simultaneously, will be used to establish a library of detected materials to identify and categorize incoming products (e.g.

Precision Phosphene Control Through Non-invasive Cutaneous Stimulation

Through safely, painlessly, and non-invasively electrically shocking the facial skin, flashes of perceived light
(called phosphenes) can be induced into one’s visual a field. This phenomenon being fully electronically
controllable and reproducible, can be used to communicate visual information to a blind person. The intended
device will use sensors such as cameras to observe the user’s surrounding and then communicate the observed
to the user in the form of phosphenes.

Spinning Multi-Beam LiDAR (VLP-16) new mapping scheme and the effect on the generated 3D point cloud : Point density and Thin features extraction in a Mobile mode of operation

Maps are vital in our life. Three-dimensional (3D) maps are essential in traditional and new applications, such as smart cities, autonomous vehicles and augmented reality. The number of end-users who require 3D maps has expanded exponentially in recent years and is anticipated to expand even more in the future. LiDAR scanners are the main sensors in 3D mapping systems. The commercially available 3D LiDAR-based mapping systems tend to be bulky, expensive and thus out of the reach of many end-users. Recently a relatively low-cost LiDAR scanner has been introduced for autonomous vehicles.

Development of Energy Flow Emulators for Greenhouse and Industrial Facilities

360 Energy is a Canadian company specializing in energy management, providing consultation services for the, commercial, greenhouses, and industrial sectors. Their current workflow utilizes a proprietary tool that that is labour intensive, which this research will seek to automate. The proposed approach is to create a building model that will output an energy breakdown of the building. The interns will undergo research to quantify the key variables for their building, exploring various data capture techniques utilizing sensors, and algorithms for statistical analysis.

Development of a biocompatible coating for an implantable magnetic marker

Non-palpable breast cancers, needing localization before removal, make approximately 60% of the diagnosed breast cancers. The gold standard for localizing includes implanting protruding wires or radioactive seeds in the lesion. The wires are not accurate, cause pain and discomfort to the patients, while the radioactive seeds warrant strict administrative and safety requirements. MOLLI Surgical Inc. offers a wireless and non-radioactive alternative: Magnetic Occult Lesion Localization Instrument (MOLLI).

AI based acoustic signal processing and analysis and its applications

Artificial intelligence (AI) provides powerful tools to many acoustic related applications. This proposal aims at acoustic signal processing and analysis and its applications using AI technologies. In order for AI to make progress in understanding the acoustic world around us, it needs to be interpreted along with multimodal messages. Three subprojects, i.e., multi-modal acoustic scene recognition, audio-visual speech recognition and audio signal processing for healthcare purpose using deep learning techniques are included in this proposal.