Innovative Projects Realized

Synthetic Image Generation for Training and Testing DNN-based Anomaly Detection Systems

Deep neural networks (DNNs) are increasingly employed in vision-based systems, extending well beyond self-driving vehicles. With the rise of the Internet of Things (IoT) and distributed vision sensing, vision-based systems have found diverse applications across various domains. These include detecting hazards on work sites and anomalies in critical infrastructure, utilizing, for example, video and image data captured by drones. In collaboration with SmartInside AI, this project aims to address the identification of equipment anomalies in the power grid using video and image data. The primary challenge that this project aims to tackle is the limited availability of high-quality labeled data for developing and testing DNN-based approaches. To overcome this challenge, we will capitalize on recent advancements in generative AI to create synthetic images. This approach will enable us to adapt SmartInside AI’s existing solution, originally developed using South Korean data, to suit the Canadian environment. Specifically, we will employ generative AI techniques to modify the existing DNN-based solution, accommodating the variations in equipment types and weather conditions between South Korea and Canada…

Développement technologique d’un ioniseur ne générant pas d’ozone et fonctionnant par effet photoélectrique

L’objectif de la recherche est d’optimiser la neutralisation de particules fines et de composés toxiques volatiles dans des milieux micro pollués par la production d’un vent concentré en ions négatifs. Les validations porteront principalement sur la portée d’action (objectif de deux mètres) du vent ionique, ainsi que la rapidité et fiabilité de la neutralisation des contaminants dans l’air par les ions négatifs. Un prototype sera conçu, fabriqué et mis à l’essai dans les laboratoires de l’Université de Sherbrooke. En fonction des performances mesurées en laboratoire, des itérations seront effectuées pour améliorer les performances du prototype jusqu’à l’obtention de performances intéressantes. Une étude de la viabilité de la technologie sera ensuite réalisée.

Combustion studies of carbon-free hydrogen/ammonia blends

This project focuses on studying the combustion behaviour of hydrogen and ammonia blends as potential alternative fuels. While clean energy sources have progressed, specific industries and transportation still rely on combustion technologies that produce harmful carbon emissions and contribute to climate change. The project aims to improve the efficiency and reduce emissions of combustion processes by exploring hydrogen and ammonia blends. The project involves conducting experiments using a specialized burner and analyzing the flames generated by the blends to understand their combustion dynamics. Numerical analysis will also be performed to complement the experimental investigations. The research aims to contribute to developing sustainable and efficient fuel technologies, enhance safety measures, and optimize system performance in various energy sectors.

Production of bio-based methyl ester sulfonated surfactant from epoxidized palm oil derivatives

The global push to decarbonise our energy system and all its related processes has stirred up a great interest amongst researchers and scientists to devise ways and means of improving processes that heavily rely on fossil fuels. The use of renewable sources to make chemicals and materials has received considerable attention as this alternative approach has prospects of reducing our carbon emissions. Therefore, the objective of this research project is to contribute to the understanding and advancement of renewable materials and sustainable manufacturing processes through the development of polyurethane from renewable sources like vegetable- oils. The findings form these studies will have significant implications for various industries,including coatings and foams, by promoting the use of eco-friendly materials and reducing reliance on non-renewable resources.

Étude de l’ablation laser en régime d’absorption linéaire

L’objectif principal du projet est de mettre en évidence le lien entre l’absorption linéaire d’un matériau, qui varie avec la longueur d’onde, et l’efficacité d’ablation laser (volume/énergie) de ce dernier. L’étude proposée consistera à usiner divers matériaux, dont les spectres d’absorption ont été préalablement caractérisés lors de projet antérieur, à l’aide de diverses sources laser notamment un laser CO2 émettant à une longueur d’onde de 10.6 ?m, ainsi que divers lasers à fibre optique émettant à des longueurs d’onde discrètes comprises entre 1 ?m et 3.5 ?m.

Fast optimization of time dependent vehicle routing problems with profits

In this project we will study distribution problems arising when time-sensitive information needs to be collected
and handled appropriately. These problems arise as a business need of our partner organization to support the
decision systems of several customers that require fast and reliable optimization methods to design distribution
plans under tight time considerations. Either the business hours of demand points or their preferences for
inventory replenishment decisions are a crucial part of the problems under interest. This internship will focus on
understanding, modeling and proposing potential avenues of research for handling and integrating the different
complex attributes arising in these distribution problems.

Paramétrisation de stratégies de Gestion de la reprise après panne

Suite à une panne du réseau électrique, les principaux équipements des ménages voient leur consommation énergétique s’accroître pour combler le déficit accumulé pendant la panne. Par exemple, les équipements de chauffage et de climatisation s’activeront à pleine puissance jusqu’au rétablissement des températures de consigne. Il en résulte, au niveau du réseau de distribution, une demande supérieure à celle rencontrée en période de pointe. C’est donc la puissance en période de reprise qui est utilisée aux fins du dimensionnement du réseau de distribution. Si on limitait la puissance associée à la reprise, on pourrait réduire les besoins d’investissements qui seront associés à l’accroissement général de la demande en électricité, qui sera substantielle si on veut atteindre les objectifs de décarbonation de l’économie.
Ce projet vise la coordination des stratégies de contrôle en reprise après panne des principales charges résidentielles (chauffage, chauffe-eau, véhicule électrique) afin qu’elles résultent en un comportement agrégé plus favorable tout en demeurant acceptable pour les consommateurs. Les résultats pourraient alimenter la préparation de nouvelles normes de performance pour ces appareils.

Redesigning Transformer Neural Networks for Data-Efficient Breast Cancer Diagnosis

Transformer neural network models have achieved state of the art results in many areas of machine learning, including natural language processing (e.g. BERT, GPT-3) and computer vision (e.g. ViT, SAM). However, to achieve these results currently requires extremely large training datasets which are not available in every domain. One such domain is medical imaging, in which data is often limited due to privacy concerns and costs. This project aims to overcome this limitation by integrating domain-specific knowledge into a transformer pre-trained on large general datasets. In contrast to previous work, which generally treats the transformer as a black-box, our approach will integrate domain-specific knowledge by modifying the transformer architecture.

Vers une vitrine technologique connectée 5.0

Les vitrines technologiques sont des outils précieux pour les entreprises, offrant des solutions innovantes qui améliorent leur performance et leur compétitivité. Elles favorisent l’innovation en exposant des avancées technologiques de pointe. En adoptant ces nouvelles technologies, les entreprises peuvent augmenter leur productivité, réduire les coûts et améliorer la qualité de leurs produits ou services. Les vitrines technologiques créent également des opportunités de collaboration et de partenariats stratégiques, renforçant ainsi la compétitivité collective de l’industrie. De plus, elles permettent la diffusion des connaissances et des bonnes pratiques, favorisant l’apprentissage mutuel au sein de l’industrie.

Ce projet à pour but de concevoir, déployer et assurer la maintenance continue d’un espace innovant qui exploite les avancées technologiques pour optimiser les processus opérationnels, à accroître l’efficacité, améliorer la performance et la compétitivité des entreprises et à favoriser l’innovation au sein de l’industrie. Il sera question de mettre en œuvre des solutions connectées, telles que l’Internet des objets, l’intelligence artificielle et la réalité augmentée, pour offrir une expérience utilisateur immersive et personnalisée.

Translating Advanced 3D Ultrasound System for Measuring Dysplasia Metrics into Clinical Practice

In this work we seek to deploy a web app to help orthopaedic clinicians all around the world analyse 3D ultrasound (US) volumes of hips to more accurately diagnose developmental dysplasia of the hip in infants. We intend to
develop and evaluate artificial intelligence (AI) models capable of segmenting relevant hip bones from 3D US volume and subsequently calculate geometric metrics indicative of dysplasia. Our primary focus is to develop
novel AI training protocols that enable better generalization and thus ease the deployment of the models to the clinical setting. Specifically, we intend to design methods to augment the data with which the AI models are trained
in ways that are congruent with ultrasound imaging acquisition. Once we evaluate the capabilities of our AI model and address potential translational limitations, our partner will be able to deploy the AI model in the clinical
environment and benefit from advanced analysis techniques that could help better understand and diagnose this condition.

Investigation of Sargassum-Based Bioethanol Blends as a Sustainable Alternative Fuel

This project seeks to contribute to the global cardon emissions reduction agenda through the use of alternative sources of fuel that do not produce carbon emissions. In other words, this approach is known as “CO2 avoidance” through the use of carbon-neutral fuels. The production of biofuel, namely bioethanol using edible feedstock has raised serious humanity concerns along the lines of the issue of malnutrition and hunger prevalent in our day. This has shifted the focus to using non-edible feedstock such as sea-weed (Sargassum) to produce bioethanol and its blends. Sea weeds offers an excellent alternative to displace fossil fuels and from an ecological point of view, sea weeds aid in reducing the carbon dioxide in the atmosphere and supplying oxygen :to the sea. For this reason, sea weed has been proposed as a promising biomass material that can be easily converted to bioethanol based on its low lignin (an unwanted material in the biomass to ethanol conversion process) content.

Pathfinder MVP 2.0 Development Project

Pathfinder365 is a digital community marketplace that focuses on the tourism sector. By harnessing the collective benefits of an interactive two sided marketplace, a map interface, a dynamic suite of features and tools, and an algorithm-free environment, Pathfinder is positioned to become the most influential digital aggregator for the micro-tourism industry. Pathfinder is in its second year of development and is currently undergoing a full market launch and establishing traction in select provinces. It’s also initiating a dynamic sales and marketing strategy.

While Pathfinder is working with an agency team, it’s also bringing developers onto its own internal team, who will work in concert with the agency’s developers. That way Pathfinder establishes its internal team, while benefiting from the stabilizing effect and experience of a seasoned professional agency.

Pathfinder has one awesome developer starting in April on an IRAP Youth Grant, and it has a second highly skilled developer, an international MBA student at the U of S, who’s currently working under the agency, but Pathfinder is looking to migrate him onto a grant program like the Mitacs BSI. Both of these developers will be absolutely essential to moving Pathfinder from the MVP stage to establishing product market fit, and will be the foundation of helping Pathfinder effectively scale throughout Saskatchewan, Manitoba, Alberta and British Columbia.