Projets novateurs réalisés

L2M – Smart Concrete for Structural Health Monitoring (SCSHM)

This project explores the development and application of smart concrete an innovative construction material enhanced with advanced sensing and self-monitoring capabilities. Unlike conventional concrete, smart concrete integrates conductive or functional additives that enable it to detect stress, strain, cracks, and environmental changes in real time. This technology addresses critical global challenges by improving infrastructure durability, reducing maintenance costs, and enhancing public safety. Our research focuses on optimizing material composition, evaluating mechanical and electrical properties, and testing performance under various environmental and structural conditions. By combining sustainability with intelligent functionality, this project aims to create greener, longer-lasting, and safer infrastructure solutions while paving the way for future smart cities and environmentally responsible construction practices.

L2M-Designing a Stable CD19 Protein Mimic for CAR T-cell Therapy

CAR T-cell therapy is a type of immunotherapy that displays remarkable potential to treat B-cell cancers. In this treatment, a protein on the surface of B-cells called CD19 is targeted, allowing for the specific killing of B-cells in patients. However, CD19 is a “difficult to express” protein, characterized by high levels of aggregated protein and low yields of active, folded protein.3 Furthermore, CD19 must currently be expressed and purified using mammalian cell lines which are expensive and time consuming. Overall, the use of CD19 for analyzing and testing CAR T-cell cultures is a major bottleneck in the developmental pipeline, and these inefficiencies greatly hinder the process of working with CAR T-cells. In this project, we have designed a CD19-like protein mimic that allows researchers to streamline CD19 production for CAR T-cell therapies by cutting down on costs and time currently required in this process.

L2M- Potatoleaf Doctor( AI model for potato leaf detection)

Agriculture supports food security for 60% of the global population, with potatoes, the 4th most consumed crop, feeding around 1.5 billion people daily (Afakh et al. and LeCun et al.). Canada produces about 5.7 million tonnes of potatoes each year, making it the 12th-largest producer in the world. The country’s potato exports generate around $1.6 billion annually. However, Canadian potato producers lose a significant amount of yield each year due to preventable leaf diseases such as Late Blight and Early Blight. Most small and mid-scale farmers still rely on manual scouting, which is slow, subjective, and often too late to prevent damage. To solve this problem, we plan to design a lightweight AI framework (using computer vision technology) that identifies leaf diseases directly from smartphone/drone images, which also works offline, providing instant and affordable diagnostics. We believe that this model will work at the early stage of detection and contribute to potato growth.

L2M – SmartRail Hotspot Analytics Using Fiber Sensors

The proposed project will use existing telecom fiber installed alongside railways to monitor the physical condition of Canada’s aging rail network in real time. By attaching small Fiber Bragg Grating (FBG) sensors at key stress points—such as bridges, track curves, and transitions—the system will continuously measure strain, temperature, and wheel impacts without adding new cabling. This information will help the partner organization detect potential problems like rail buckling, cracking, or settlement early, so they can fix issues before service disruptions occur. The expected benefit is improved safety, fewer costly delays, and more efficient use of maintenance resources, extending the life of Canada’s rail assets while ensuring reliable freight and passenger movement.

L2M-Augmented Reality-Based Training System using Multimodal Language Model for Context-Aware Guidance and Activity Recognition in Complex Machine Operations

Many industrial companies still rely on traditional training methods and struggle to keep up with evolving skill requirements. These conventional approaches, such as manuals, videos, and classroom instruction, are ineffective in delivering the hands-on skills required to operate complex machinery.
Our project introduces an Augmented Reality (AR) based training platform powered by Multi-Large Language Models (MLLMs) that acts as an intelligent instructor. It can understand what the user is doing, read the machine’s feedback, and guide the user directly on the equipment. Unlike existing AR systems that follow a fixed, pre-scripted path, our system continuously adapts to user actions and updates the instructions automatically, allowing trainees to learn safely and independently without constant supervision.
The system integrates structured prompt design, model-target detection, and MLLM-based reasoning to interpret visual and textual cues in real time. It can also be easily customized for different machine types and industry-specific workflows, enabling rapid deployment across diverse applications. These capabilities position the solution as technically innovative and practically scalable, bridging the gap between AI-driven research and real-world machine-operation training.

L2M-Veterinary point-of-care test device to apply novel immune marker for improved pet oral gum health

Periodontal disease (PD) is a chronic bacterial infection of the gumline and tissues that anchor teeth and according to the Canadian Dental Association (CDA) 7 out of 10 Canadians will experience some form of this disease over their lifetime. PD is caused by specific pathogenic anaerobic (those that prefer low oxygen environments) bacterial species known as the “Red Complex” that establish chronic infection as a sticky difficult to remove biofilm. The “Red Complex” bacteria use sugars for energy differently than the healthy oral bacteria and unique byproducts that they produce provide targets for diagnostic testing to detect PD much earlier than current physical examination methods. In addition to tooth loss, pathogenic bacteria responsible for PD are also linked to increased health risks including cancer, diabetes, heart infections and even degenerative diseases such as Alzheimer’s. Similar to humans, companion veterinary animals more than 3 years of age such as canines (dogs) and felines (cats), experience increased incidence of PD, which is considered one of the top factors impacting quality of life for animals as they age. Diagnosing PD in companion animals is also more complicated than in humans and often involves dangerous and costly anesthetization to perform imaging techniques such as X-ray. The treatments and cleanings to address PD in these animals is extremely expensive making it prohibitive for many pet owners delaying care until it is too late. Having a cost-effective and rapid test for veterinary oral health would allow earlier detection of PD and allow for interventions that would prevent future high-cost and high-risk treatments. This novel technology could also be adapted to larger agricultural animals having positive impacts on Canadian animal health and the economy. The valuable network contacts and resources provided by the partner organization will assist with market analysis and IP strategies for this novel invention.

L2M – TerraFire Intelligence Platform: An Integrated Solution for Predictive Wildfire Intelligence

Wildfires create serious threats to communities, infrastructure, and natural environments across Canada. Many regions face fires that grow faster and burn hotter due to climate change, dry conditions, and limited access to real-time information. Current tools often fail to provide timely or accurate understandings, especially in remote or northern areas where ground reports arrive late and aerial surveys are too costly. As a result, emergency teams and infrastructure operators face significant challenges in detecting fire activity early, predicting danger zones, and protecting people and assets.
This project advances TerraXAI Wildfire Intelligence, a Canadian platform that integrates satellite data, artificial intelligence, and ground-based sensor measurements into a unified system. TerraXAI provides real-time understanding of ignition risk, fire spread, and exposure to critical infrastructure. The platform draws from multi-source satellite images, weather data, and readings from sensors that monitor greenhouse gases, temperature, humidity, and smoke. It then produces clear maps and alerts that help decision-makers understand fire conditions with far greater accuracy.
The project’s goal is to create the technical components needed for an operational version of TerraXAI. This includes a unified data pipeline, refined AI models, and a WebGIS interface that presents intuitive, easy-to-use visuals. The system aims to support wildfire teams, utility operators, insurers, and local governments as they plan responses, reduce losses, and strengthen community safety.
This work clearly offers a public benefit. A reliable early-warning and prediction tool reduces emergency costs, limits damage to forests and infrastructure, and supports climate-resilience planning. TerraXAI also supports Indigenous and rural communities that face high wildfire exposure but often lack fast access to fire intelligence. By developing this platform in Canada, the project enhances national capacity in environmental monitoring, advanced analytics, and climate adaptation technology.

L2M – Polycoat Flooring

The proposed technology, called polycoat flooring, aims to ease the electricity bill of Canadian households through no added effort, simply by harnessing otherwise wasted mechanical energy and transforming it into usable electricity. Through a simple and highly intuitive installation process, this product provides an alternative aesthetically pleasing flooring option that generates electricity by simply walking on its surface. This electricity can be used instantly by charging small electronics for example, or it can be stored for later use. Bridging the gap between research and the market for technologies such as the polycoat flooring is the main driving force behind Springboard Atlantic’s core mission, and by supporting this project, they will help drive economic growth in Nova Scotia, and make an impact where it is most needed for Canadians.

L2M – WayLuma

This project will develop an easy-to-use system that helps students automatically fill out university application forms by organizing their information and entering it into different school portals. The goal is to save students time, reduce mistakes, and make the application process less stressful. For North Forge, this project provides a chance to support a new technology that can help thousands of students and families, while also giving them early access to an innovative tool being built through the Lab2Market Validate program. This partnership also allows North Forge to strengthen its role in supporting youth-focused and education-focused startups in Manitoba.

L2M – Market Validation and Commercialization Assessment of Sustainable Activated Carbon Produced Using a Novel Production Method

This project focuses on validating the market potential and commercialization pathway for a sustainable activated carbon product produced using novel pyrolysis of Brewer’s Spent Grain (brewery waste). The work will assess customer segments, value propositions, production feasibility, cost structure, and pilot-scale readiness for Maple Carbon Technologies Inc. The internship will generate evidence-based insights to support market entry, identify early adopters, evaluate competitive positioning, and define the commercialization strategy for launching bio-based activated carbon within environmental, water treatment, and sustainability sectors.

Solving the Distributionally Robust Split Delivery Vehicle Routing Problem via exact methods

This Mitacs Globalink project, part of the PhD thesis titled “Robust Optimization for Logistic Problems,” aims to develop new optimization techniques for solving the distributionally robust split delivery vehicle routing problem under uncertainty. In split delivery routing, customer demands can be served in multiple partial shipments rather than a single delivery, an approach that can reduce total delivery costs by up to 50%, though it significantly increases computational complexity to find the optimal solution. This project will study the problem through the distributionally robust optimization framework, which allows the incorporation of limited probabilistic information without requiring perfect knowledge of the underlying distribution of parameters, a common limitation in real-world applications. As this variant has not yet been explored in the literature, it holds strong potential for improving the efficiency and resilience of modern logistics and delivery systems. Building on the applicant’s previous successful implementations of branch-and-cut algorithms for deterministic and robust variants of the problem, this project seeks to develop an exact method for the new distributionally robust framework. The collaboration between experts from Université Laval and the University of Groningen combines expertise in transportation optimization and stochastic and distributionally robust programming, a promising outcomes that will advance both domains.

IA générative et intelligence éducative : vers des évaluations intégratives multimodales

Les systèmes tutoriels intelligents sont de plus en plus utilisés pour offrir un apprentissage personnalisé dans les environnements numériques. Ils permettent d’adapter les activités aux besoins de chaque apprenant, de suivre sa progression en temps réel et de fournir un accompagnement individualisé. Ces technologies génèrent une grande quantité de données sur la façon dont les étudiants apprennent, par exemple leur niveau d’engagement, leurs stratégies d’étude ou leurs difficultés.

Cependant, les méthodes actuelles d’évaluation utilisent encore des mesures ponctuelles, comme des tests ou des notes finales, qui reflètent mal la progression globale des apprenants et leurs compétences complexes, telles que la pensée critique ou l’autonomie. Bien que l’intelligence artificielle générative puisse analyser des réponses écrites et offrir du feedback personnalisé, elle n’est pas encore pleinement intégrée dans des méthodes d’évaluation complètes et cohérentes.

Ce projet vise à développer une nouvelle approche d’évaluation intégrative, assistée par l’IA générative, capable de combiner différentes sources d’information sur l’apprentissage. En analysant à la fois les comportements d’étude, les productions écrites et les résultats, le système pourra produire des diagnostics clairs et personnalisés. Cette approche aidera étudiants et enseignants à mieux comprendre les progrès, identifier les difficultés et soutenir un apprentissage plus efficace et équitable.