Innovative Projects Realized

FeatTS-Detect : Features-Driven Time Series Anomaly Detection

With the rapid growth of sensors in Cyber-Physical Systems such as clinical data, industrial systems and data centres, there is an increasing need to monitor these devices to secure them against anomalies. This is particularly the case for streaming clinical data. Indeed, the timeliness revelation of anomalies in these data can save the patient’s life.
Time series anomaly detection has been a perennially important topic in data science, but in recent years there has been an explosion of interest in this topic.
The main objective of the project is to implement a system that permits to discover anomaly points and the anomaly subsequence of points in the time series data through a system that evaluates an anomaly considering a set of features extracted among the time series. The features have obtained good performance for clustering of time series. Therefore, our idea is to extend this technique for detecting anomalies.

Mass spectrometry imaging of glucocorticoids in the mouse thymus

The project aims to better understand how stress hormones (i.e., glucocorticoids) are synthesized in the thymus, a key primary lymphoid organ of the immune system. Stress hormones are critical for proper thymic development and immune function. However, many details about how and where in the thymus stress hormones are produced remain unexplored. As part of the proposed project, the student will be visiting the University of Edinburgh to access state-of-the-art mass spectrometry (MS) imaging. MS imaginng is a revolutionary technique that permits regional analysis of stress hormones directly from organ tissue surfaces. No research institution in North America possesses MS imaging for glucocorticoid detection, and the host supervisor (Prof. Ruth Andrew) has the unique expertise required. This project will produce a “map” of stress hormone distribution and levels within the thymus using a mouse model. The student will fill critical gaps in knowledge in immune-endocrine crosstalk that can be applied to interventions and preventions in immune system disorders.

Human-in-the-loop: Occupants as integral drivers of indoor climate controls

Occupant thermal comfort and ventilation drive the operation of HVAC (heating, ventilation, and air conditioning) systems in buildings, which consume a large portion of their energy use. “Human-in-the-loop” (HITL), a term borrowed from machine learning referring to a synergy between humans and machines, is a data-driven approach that aims to enable human-based controls of the HVAC system. This participatory approach integrates environmental data and recurrent occupant feedback in the HVAC control loop to tune comfort predictions and determine set points. The goal of this project is to apply it to an educational setting, to attempt to optimize educational building environments for teaching and learning, while minimizing energy use.
Following a machine-leaning based approach, this research aims to develop a pilot testbed project in the BCIT campus to explore the application of human-in-the-loop principles in an educational setting. The objective of the project is to design an experiment in a classroom or a group of classrooms linking a comfort App in the students’ and faculty phones, with enhanced room environmental sensors to provide real time feedback to the HVAC building management system (BMS) controls.

AI Orchestrator For Energy-Efficient Future Networks

The information and communication technologies (ICT) sector is an energy intensive and growing sector (with an increase of 9% annually). Recently introduced applications of the next-generation radio networks, namely 5G-and-beyond (5GB), have dramatically accelerated the use of ICT services in many economic sectors, creating a unique opportunity to improve our quality of life. In this project, we will optimize a cloud-edge fabric model with respect to new requirements of next-generation radio applications and minimized energy consumption. Based on an artificial intelligence (AI) orchestrator, this model will be able to automatically adapt to the requirements of different classes of users while being eco-responsible. Services will be provided to end users through virtual network slices, which are optimized from end to end. The findings of the project will be assessed and quantified on the ENCQOR (Evolution of Networked Services through a Corridor in Quebec and Ontario for Research and Innovation) network. This outcome will have the potential to be standardized in OpenRAN (O-RAN) alliance and hence be used worldwide.

Transactive Energy Control Mechanism for Future Power System

The growth of renewable energies on the demand side, as an important step for decarbonization and transition to the smart grid, poses serious challenges to control the system and maintain stability. Transactive energy control (TEC) is introduced as an effective approach in which distributed control methods are utilized rather than centralized ones. Although many studies have been devoted to utilizing demand flexibility, electric vehicle, and microgrids for control purposes in TEC scheme, there is a far distance to implement these studies in reality. The reasons are ignoring or partially considering uncertainty management, connect networks, the relation among transmission systems and distribution systems. To tackle these challenges, a TEC scheme is proposed the proposal in which uncertainty is managed locally and all agents (i.e., power plants, big demands, distribution systems, aggregators, microgrids, etc.) are hierarchically connected to each other. Furthermore, several separate networks are considered according to reality. To implement this scheme alternating direction method of multipliers (ADMM) will be utilized.

Improving neural interface selectivity by measuring evoked neurophysiological responses

Electrical stimulation of nerves can be used to restore movements to individuals who are paralyzed following spinal cord injury. Our project is divided into two main parts. The first part aims to combine measurements of muscle and nerve activity evoked by electrical stimulation in order to improve our understanding of the mechanisms involved in nerve stimulation using electrodes composed of several contacts. The second part consists of modelling the influence of the choice of different stimulation configurations (current distributions within the electrode contacts) on the activation of nerve fascicles (subsets of neurons grouped in the nerve). The results of this work will help to create technology that can restore more precise and effective movements after paralysis.

Mental Health and Peer Support Among Adolescents Living with Disability or Chronic Illness in Canada and France

Two students will be involved in this project each visiting the other’s country and undertaking collaborative research in a Public Health setting. We are requesting funds to cover the costs for the Canadian student to travel to France, costs for the French student will be covered from elsewhere. The students will join national level research teams already established as part of the World Health Organization supported Health Behaviour in School-aged Children (HBSC) Study. We are interested in exploring mental health and peer support for early adolescents (aged 10-16 years) living with and without disability or chronic illness in Canada and France. These two countries identify “disability status”” in different ways which is also interesting from a measurement perspective. Students will undertake statistical analysis using HBSC data from both countries under the guidance of the identified supervisors. The French student will be situated at Queen’s University and will work with Canadian data. The Canadian student (funded through MITACS) will be at the École des Hautes Etudes en Santé Publique and will work with French HBSC data. This study will help inform HBSC international teams from across 50 countries who make decisions about HBSC survey items every four years.”

An application of lignin-based coating to structural natural fiber composites

A Canadian-based company, Terrasol Geosolar Inc., has identified a unique opportunity to offer hurricane-resistant solar energy infrastructure for weather-ravaged areas of the world. Towards this functionality, there has also been significant market incentives to include bio-derived composite materials within the structure. However, current developments of natural fibre reinforced composites (NFRCs) in structural applications has been limited by several factors, including the weak adhesion between the polymer matrices (often hydrophobic) and the natural fibres (hydrophilic). This joint research proposal via the MITACS internship program aims to investigate the use of lignin to improve the water-resistance of natural fibres reinforcing biobased polymers in the design of hurricane-resistant solar panels structures. The lignin compatibilization effect will be analyzed by adhesion forces measurement between polymer and coated fibres at nanoscale and microscale. The effect of the lignin coating on the mechanical, chemical and thermal properties of the samples will also be investigated.

Caractérisation de scintillateurs hétérogènes pour radiographie et tomodensitométrie par temps de vol de photons

L’imagerie utilisant les rayons X comme la radiographie et la tomodensitométrie sont entachées par un phénomène physique appelé la diffusion. Cette dernière provient du fait que les photons dévient de leur trajectoire originale et sont détectés dans des pixels voisins voire sont déviés hors de la caméra et ne sont pas détectés. Bien que l’on puisse utiliser des grilles anti-diffusantes ou encore des techniques de traitement de signal post-acquisition pour mitiger le problème, il est encore nécessaire d’augmenter la dose pour palier ces problèmes. L’Université de Sherbrooke propose d’utiliser le temps de vol de photons pour discriminer les photons ayant voyagé en ligne droite des photons ayant diffusé et mis plus de temps pour être détecté. Le principe repose sur l’utilisation d’une source rayon X pulsée et de photodétecteurs résolus en temps. Un scintillateur, capable d’arrêter les rayons-X et de transformer cette énergie en photons lumineux, est déposé sur le photodétecteur. Le stage consistera à caractériser et évaluer la performance de différents scintillateurs (LSO, LYSO, plastique et matériaux quantiques) ainsi que différents empilements de ces détecteurs pour déterminer la meilleure combinaison en vue d’apposer une estampe de temps à chaque rayon-X détecté.

Detection of anomalous emotional responses using attention mechanisms for deep machine learning

Computer-based multimodal affect recognition methods fuse multiple informational channels, typically video, audio, and text, to resolve the emotional state of a monitored individual.
The proposed research aims to develop multimodal deep learning models to recognize anomalous emotional responses, which correspond to a deviation from the expected affective reaction for a particular context. Since multimodal affect recognition applications involve large inputs such as video and audio frames and text passages, training deep neural networks to recognize anomaly presents significant challenges as the model may be unable to optimally maintain the spatial and sequential information. Hence, we propose to employ attention mechanisms to increase the emphasis on relevant spatial and temporal relationships in the input data. Attention mechanisms have been mainly applied for natural language processing applications; however, we hypothesize that we can develop analogous approaches for video and audio signals to improve model performance.

AI based technology adoption in circular economics

This project is a cross-disciplinary study of econometrics and machine learning (ML) models applied to the decision making modelling in industry. The problematic arises from the lack of tools supporting the transition to circular economics model and the need to identify the key factors to influence this transition.

The project aims to explore the key elements affecting the high management decision making process in technology adoption. A discrete choice experiment survey will be conducted. The data analysis procedure will involve both econometrics and machine learning techniques. Canadian partners will provide the information sources in industrial domain, as well as the knowledge in advanced ML and AI modelling techniques. French side posesses all the required knowledge and skills for analysis of human behaviour. The simulation and theory-testing framework proposed in previous works, will constitute the core of modelling approach, allowing to increase the reliability of results.

Exploring the boundaries of serial electron diffraction

Serial electron diffraction (serial-ED) crystallography is an emerging structural biology method whereby data are collected from protein nanocrystals using a sub-micron-sized electron beam. The method is set to lead the future of nanocrystallography, given its cost-effectiveness and the small amount of sample required. Combining the expertise of the Colletier and Miller teams, the project will address three issues of fundamental importance to further advance the serial-ED methodology. Specifically, we will (i) test if the method can be applied to solve nano-crystalline protein structures in the cellular environment (i.e. in vivo); (ii) determine what is the smallest crystal size that can be probed by serial-ED; and (iii) offer a proof of feasibility for time-resolved serial-ED experiments on macromolecular nanocrystals. Results from our Serial-ED-boundaries project will allow to explore the boundaries of serial-ED and to develop its full promise for nano-crystallography.