Projets novateurs réalisés

Optimal Energy Management in Urban Microgrids using Reinforcement Learning

This research project addresses the critical challenge of managing energy in residential urban areas, where the rapid adoption of rooftop solar panels and electric vehicles is placing unprecedented pressure on local power grids. Instead of relying on traditional forecasting methods that struggle with the unpredictable nature of household energy usage, this project develops an intelligent Energy Management System using Reinforcement Learning. This AI agent learns to coordinate residential devices in real-time, optimizing when to store excess solar energy in batteries and proactively managing EV charging to support the grid rather than strain it. The project benefits the participating institutions by combining Canadian expertise in advanced urban building modeling with Vietnam’s practical insights into volatile grid operations, resulting in a novel, transferable control methodology that strengthens international research collaboration and future smart city innovation.

Hunting Adversaries in Federated IDS: Does Local Adversarial Training Actually Make the Global Detector Harder to Evade?

The objective of the project is to enhance the resilience of Federated Intrusion Detection Systems (FIDS) against adversarial attacks. Local Adversarial Training (LAT) will be used so that every client can develop and train on weak adversarial traffic in the local environment to enhance the robustness of the global model against potential poisoning and evasion attempts. The evaluation of this defense will include realistic federated contexts, including non-IID data and models undermined by a Byzantine participant. The project will also incorporate two dimensions into the evaluation of the system by combining machine learning robustness metrics and operational cybersecurity metrics—such as MITRE ATT&CK mappings, model drift detection, and time-to-detection—to better indicate how the AI robustness will operate in a real-world SOC context. The findings will result in a concrete basis for the development of privacy-preserving, trustworthy intrusion detection systems, while also contributing to the state-of-the-art in trust architectures for AI and resilience of Canada’s digital infrastructure.

AI Driven Cooperative Learning and Intelligence for Collision Prediction in Autonomous Vehicle

Autonomous vehicles are becoming more common, but they still face real challenges when driving in busy or unpredictable environments. This project focuses on creating a smarter and safer way for these vehicles to learn how to avoid collisions and make better decisions on the road. The idea is to teach the system from expert driving examples, allow different vehicles to learn from one another without sharing private data, and make the final model lightweight enough to run on small onboard computers. The work will be tested both in simulation and on the Quanser QCar platform at Concordia University, ensuring that the results are practical and ready for real-world use. This collaboration will strengthen research ties between both institutions and contribute to Canada’s growing leadership in intelligent transportation. In the end, the project aims to help researchers develop safer, more adaptable autonomous vehicles that can handle a wider range of real-world conditions.

Automated Customized 3D-Printed Wrist Splints from High-Fidelity Low-Cost Photogrammetry Scans

This project creates a fast, low-cost pipeline for making custom wrist splints using photogrammetry, automatic CAD modeling, and 3D printing. A patient places their arm in a compact scanning rig, and a full 3D model is captured instantly using multiple synchronized cameras. Software then generates a lightweight, ventilated splint design automatically from a few anatomical landmarks. The splint is printed on-site so the patient can receive it within about one hour of scanning.

The final outcome is an open, reusable workflow (including scripts, CAD templates, and printing settings) that partner institutions can use for teaching, research, and rapid clinical prototyping. By automating the design and fabrication steps, the system shortens wait times and frees hospital staff from routine orthosis production.

IVP-DES – Vitalité et relations interpersonnelles : mesurer le bien-être des directions d’établissement scolaire

Les directions d’établissement scolaire (DES) contribuent à créer un cadre propice à la réussite des élèves, en exerçant notamment une influence sur le climat scolaire ainsi que la motivation du personnel (Grissom, Egalite, & Lindsay, 2021). Afin d’accomplir leurs missions, celles-ci doivent continuer à développer un ensemble de compétences à la fois professionnelles, émotionnelles et sociales (Hadchiti, Frenette, Dussault & Loye, 2024). Ces dernières agissant notamment sur la qualité de vie au travail. Or, les travaux de recherche témoignent d’une persistance des risques psycho-sociaux dans le vécu du travail (Gravelle et al., 2023). À ce titre, ce projet de recherche s’intéresse au bien-être des directions d’établissement scolaire à l’aune de deux indicateurs que sont la qualité des relations interpersonnelles ainsi que les compétences psychosociales. Plus spécifiquement, il aspire à construire un Indice de Vitalité Professionnelle (IVP-DES) permettant de mesurer le bien-être des directions d’établissement, en mettant l’accent sur ces deux indicateurs. Une perspective de la recherche encore peu explorée dans les travaux, et dont l’enjeu, à terme, est de fournir un outil de mesure efficace et applicable aux milieux concernés, soit les directions d’établissement scolaire, dans une visée d’améliorer le bien-être au travail.

Clean Energy and the Sharing of Traditional Indigenous Knowledge

The Petroleum Technology Research Centre, in league with the First Nations Power Authority, has been developing awareness of clean energy options for most of the First Nations Communities located in Canada, with a particular focus on Alberta and Saskatchewan. Over the past 18 months this has included a number of forums, one-on-one discussions, and smaller focused conferences and surveys to determine some of the best options tied to the unique situations, geographical locations, and geology of those communities. This project, growing out of the work PTRC has done in collaboration with the FNPA, will utilize and train an undergraduate intern (Dakota Smoker) in a new capacity as forum planner, content developer, and liaison with different Indigenous communities and non-indigenous partners like universities to increase awareness of traditional Indigenous knowledge in academe and industry. These tasks will be conducted in close supervision by the senior management (CEO) of the PTRC in consultation with an academic supervisor at the First Nations University.

Automatic identification of metaparameters in machine learning and signal processing applications (with UNaM)

This project will focus on the methods and algorithms which facilitate automatic choice of the parameters required in estimation/learning from data. We will focus on real-time algorithms such as those appearing in adaptive signal processing and on-line learning. The project will be particularly focus on the identification of meta-parameters in real-time from the data while performing the estimation/learning. This is a well-known, albeit not yet solved, problem in signal processing, where meta parameters are often heuristically defined. The same issue appears in machine learning when parameters defining the convergence or regularization have to be decided prior to the estimation. We will study methods which find/estimate these parameters at the same time as the learning takes place and should become a valuable tool to fully automatize the adaptive algorithms.

Automatic identification of metaparameters in machine learning and signal processing applications (with UFSC)

This project will focus on the methods and algorithms which facilitate automatic choice of the parameters required in estimation/learning from data. We will focus on real-time algorithms such as those appearing in adaptive signal processing and on-line learning. The project will be particularly focus on the identification of meta-parameters in real-time from the data while performing the estimation/learning. This is a well-known, albeit not yet solved, problem in signal processing, where meta parameters are often heuristically defined. The same issue appears in machine learning when parameters defining the convergence or regularization have to be decided prior to the estimation. We will study methods which find/estimate these parameters at the same time as the learning takes place and should become a valuable tool to fully automatize the adaptive algorithms.

Quantum-Assisted Multisensor Fusion for Wildfire Fuel Typing & Hazard in New Brunswick

This project will build two practical wildfire tools for New Brunswick: a province-wide 10-metre map of “fuel types” (what can burn and how) and a near-real-time hazard map that shows where fire risk is rising during the season. We will combine free satellite images (Sentinel-1 radar and Sentinel-2 optical), tree-height information from space (LiDAR/GEDI), local terrain, and weather, and we will use small pieces of quantum-assisted computing to tune the models faster and keep the risk scores stable when conditions change. The maps will help fire managers, municipalities, utilities, forestry, insurers, and communities plan prevention, patrols, and response with more confidence. The University of New Brunswick will gain a working, shareable mapping pipeline and hands-on training for students in Earth observation, AI, and quantum methods; Rajiv Gandhi Institute of Petroleum Technology (India) will gain access to Canadian workflows, joint supervision and publications, and new skills that can be applied to other environmental risks, strengthening a long-term research partnership for both institutions.

Edge-Intelligent Video Systems for Elderly Support: Monitoring, Understanding, and Interactive Avatars

With global shifting toward an aging population, empowering seniors to age in place is both a social challenge and a technological opportunity. Video-based home monitoring has the great potential to provide both robust support and privacy for elderly by leveraging advanced and edge computing techniques. Working together with our industry partner Sharp Growth (Canada) Limited, this project aims to develop edge-based privacy-preserving, accurate, energy-efficient video-based AI monitoring and analysis solutions for elderly monitoring and support, by address key challenges such as limited activity understanding, lack of rich semantic feedback and summary, and limited interactive capabilities for users. The proposed work contributes to accessible health monitoring technologies, with great potential for applications such as at-home hospital.

Professional standards for the care of incarcerated people: A content analysis of organizational position statements

Women are the fastest-growing prison population in Canada, yet people in prison often face serious barriers to receiving equitable and timely healthcare. These barriers include institutional delays, lack of privacy, denial of care, and limited access to health information. Healthcare providers working in prisons can help address some of these issues, but many report feeling unprepared due to limited training, unclear policies, and uncertainty about their roles in a correctional environment. As a result, harmful aspects of the prison setting can be reproduced within healthcare encounters. To better understand these challenges, our team previously interviewed 30 clinicians who provide care to incarcerated women and gender diverse people across Canada. Participants emphasized the need for guidance and policy support from their professional colleges to ensure ethical and equitable care. Currently, no such statement or policy exists. Building on these findings, this project will systematically review existing health professional statements on the care of incarcerated people and analyze their key themes and approaches. The results will inform the creation of policy frameworks, position statements, and healthcare protocols tailored to the needs of women and gender diverse people in prisons in Canada and in comparable international contexts.

l’influence numérique chinoise en Asie du Sud-Est : le cas de l’Indonésie

Ce projet analyse comment la Chine utilise les infrastructures numériques pour exercer son pouvoir, et étendre son influence en Asie du Sud-Est, une région clé dans la rivalité sino-américaine. Il explore trois dimensions du pouvoir numérique chinois : la genèse de ce pouvoir à travers la relation d’interdépendance entre l’État chinois et les entreprises technologiques chinoises (Huawei, Alibaba) ; les conditions qui favorisent ou limitent la transformation du pouvoir numérique en véritable influence ; et le pragmatisme des États d’Asie du Sud-Est qui tirent parti de la concurrence sino-occidentale. Ce projet apporte une triple contribution. Théoriquement, il redéfinit le pouvoir numérique chinois comme une capacité à répondre aux besoins concrets des pays partenaires, plutôt qu’une simple capacité d’imposer. Empiriquement, il comble un déficit de données primaires sur les interactions entre l’État chinois, ses entreprises technologiques et les pays partenaires. Sociétalement, il éclaire les choix stratégiques des États d’Asie du Sud-Est et fournit ainsi des outils analytiques utiles aux décideurs et investisseurs occidentaux.