Innovative Projects Realized

Efficient Signal Processing and Radio Resource Management (RRM) for Coordinated Multi-node Downlink Transmission in Heterogeneous Cellular Networks

Reliable high speed wireless data transmission is the primary goal of future cellular systems. A key impediment
towards that goal is interference on receivers from transmitters in adjacent cells. Coordinating transmissions
between clusters of transmitters is one way of solving this problem. This coordination is more difficult when the
types of transmitters and cell sizes vary within a heterogeneous network (HetNet). The design of HetNets is an
emerging research field. This project has three main goals: 1) Investigate the design of coordinated HetNets. 2)
Design efficient resource allocation algorithms for these networks, for the selection and encoding of users and
their data, and the dynamic formation of clusters of transmission points. 3) Investigate how best to reduce the
feedback overhead in the sharing of information within a cluster, and the impact of this limited feedback. This
research will allow the partners to gain relevant expertise and impact the design of future wireless standards.

Co-operators – Comprendre les enjeux des communautés

Co-operators est une coopérative canadienne, chef de file de l’assurance multiproduit et des services financiers. Co-operators cherche à explorer de nouveaux modèles d’affaires pour mieux servir sa clientèle et réussir sa mission (la sécurité financière des canadiens et canadiennes) et cela nécessite des choix stratégiques, surtout au niveau des communautés suivantes : immigrants, peuples autochtones, étudiants, etc.
Le projet vise à mieux comprendre les thèmes et enjeux qui sont les plus importants pour différentes communautés cibles. Cela permettra d’extraire de l’information très importante des réseaux sociaux publics afin de mieux comprendre les besoins de la clientèle cible, une étape nécessaire au processus d’innovation.
Le but du projet est de développer un outil permettant d’extraire les thèmes principaux de discussions sur des forums publics pouvant être intégré dans une application par les ressources IT de Co-operators.

Bentley – Approche semi-supervisée pour l’identification et la localisation des bris et défectuosités

Bentley fournit des logiciels spécialisés et adaptés à tout type de projet d’infrastructure au niveau mondial. Le projet proposé vise à développer un détecteur d’anomalies générique basé sur un autoencodeur adversarial, qui permet l’apprentissage automatique de la distribution des éléments normaux. Par la suite, tout élément s’écartant de cette distribution peut être qualifié d’anormal, même si le système n’a jamais vu de tel défaut lors de son entraînement.
Globalement, l’objectif est d’assister l’expert qui analyse les données pour lui permettre de se concentrer sur les éléments cruciaux (les données indiquant des anomalies) sachant que l’identification de >95% des données ne présente aucun problème. Le projet se concentrerait d’abord sur des données liées aux égouts et aux rails, mais pourrait ensuite s’appliquer à une panoplie de situations. Bentley envisage une utilisation future conjointe de multiples modalités (visuel, infrarouge, etc.) afin d’améliorer les performances et la polyvalence de la méthode.

Développement et prise en charge d’un algorithme de synchronisation des pompes de plusieurs stations de pompage d’eaux usées

Polytechnique Montréal va développer un simulateur numérique pour recréer la réalité actuelle des stations de pompage en série dont le fonctionnement est basé sur des niveaux, ajouter un contrôle virtuel basé sur le débit et opérer les stations à distance en faisant un diagnostic en temps réel, dans le but d’effectuer une économie d’électricité du système. Ce simulateur aura accès à de vraies données de débit d’entrée, de capacité et courbe de pompe, de niveau et pression. La méthodologie proposée dans cette étude repose sur la combinaison des données de terrain, des approches théoriques et des travaux de modélisation numérique.

Creating a machine learning model to predict activity in playgrounds across North America

Biba is a company that creates interactive experiences for families on playgrounds and provides data for playground owners. This research project explores how can we leverage third party data sources and machine learning to confidently determine how many people are in a playground at a given time and how long they spent there. This kind of information is critical for park and playground stakeholders, and if we can solve this problem, Biba would be the first company to be able to produce an industry metric of this kind. Using our own app data along with various third party sources we are looking for some data scientists to plan, preprocess and train machine learning models to begin making high-confidence predictions for communities. Interns on the product will be contributing to an effectively state-of-the-art data product for the parks and recreation space.

Process Analytical Tools, Flow Cytometry and Raman Analysis for Cell Viability Studies

The objective is to develop an in-line Process Analytical Technology (PAT) tool to monitor viability and biomass levels at the fermentation stage bacterial organisms.
The Master student will be engaged in the development of the PAT tool for deployment into a state-of-the-art vaccine manufacturing facility. The student will use benchtop tools such as flow cytometry to monitor the bacterial fermentation process and develop correlations to the inline measurements. This work will be conducted as follows:
• Develop flow cytometry method for monitoring cell viability for both seed expansion and fermentation phases.
• Develop correlations with inline PAT and offline viability measurements
• Develop methodology and general protocol for the usage flow cytometry to monitor bacterial cell viability.
• Validate PAT model against in process measurements.
• Generate report summarizing work performed.

A Step Towards a Lower Carbon Future: Integrating Closed Loop Geothermal Technology in District Cooling Applications

This research aims to provide an affordable, clean energy alternative to meet the world’s cooling demands. As global warming, urbanization, and society’s dependency on digital storage increases, the world’s cooling demands continue to rapidly grow with predictions showing that they will outweigh heating demands by 2060. The majority of these demands are currently being met with the use of fossil fuels. Through the use of proprietary Eavor-Loop technology (joined horizontal wells acting like a subsurface heat exchanger), geothermal energy can become a feasible, reliable, scalable solution. This research will design an Eavor Cooling System and determine the feasibility based on process simulations, economics, available applications, and environmental benefits. The newly designed cooling system will open up a world of opportunities for companies and governments looking to transition into a lower-carbon future with a truly renewable energy that’s not plagued with intermittency, geological, or geographical problems.

AI in Ophthalmology triage automation

There are currently 18 retina specialists in the province of Alberta, approximately half in Calgary and half in Edmonton. Retinal diseases are common. For example, approximately 6.5 percent of people age 40 and older have some degree of macular degeneration. Diabetes retinopathy affects approximately 500,000 Canadians. Many retinal conditions are treatable when detected early, however retinal specialists are concentrated in large urban centres. There are 700+ optometrists across the province today capable of taking and transmitting retinal images for consultation with a retinal specialist. Consultations have significant wait times and the current consultation processes are suboptimal, in terms of ease of triage and prioritization of patients.

Calgary Retina Consultants (CRC) is the largest retina practice in Alberta and among the most respected in Canada. CRC has established a platform to receive retinal images directly from optometrists, as part of the referral process. CRC has accumulated 6000+ low-resolution 2D images from historical referrals. CRC would like to develop a process to automate identification of the diagnosis and severity from images in referrals sent from optometrists, in order to accelerate triage and workflows, and to decrease wait times for patients who would benefit from treatment.

The effect of shot peening and anodizing on the fatigue life of highstrength aluminum alloys

High-strength aluminum alloys (AA) are extensively used for landing gears given their
low density compared to steel. All AA parts undergo special processes after
machining such as shot peening and anodizing. Because the impact of the special
processes on the fatigue properties of the material is not very well known,
conservative security factors are used leading to an increase in the wall thickness of
the parts and, thus, weight increase. The objective of this project is to understand, at
the material level, the impact of special processes on the fatigue life of high strength
aluminum alloys by conducting fatigue testing, observations and analysis of the
fracture surfaces of fatigue specimens. as well near surface mechanical and
microstructure characterization. This project will help better understand the effect of
process parameters on the fatigue life of the components and will ultimately lead to
design optimization of landing gear parts.

Low-order Model of Supersonic Fluidic Oscillator for Superplastic Forming

The Superplastic Forming process involves gas injection from a variable pressure supply, to form a heated metallic sheet into a complex automotive body panel shape onto the surface of a die. The current process involves excessive forming times which allow residual stresses to relax and avoid cracking and tearing. Research shows that pressure pulsations of the gas supply increases allowable material strain rate, reducing required manufacturing time. Our research shows that a Supersonic Fluidic Oscillator, due to the absence of moving parts, is capable of reliably generating the required pressure fluctuations under the extremely high temperatures present. Design of these devices requires use of computationally expensive fluid dynamics solutions of the complex compressible flow which are prohibitive for use in industry. The objective is to develop a simplified mathematical model with suitable accuracy to allow the partner to quickly design these devices and become a world leader in this area.

Building an active learning module to accelerate the validation and qualification of image-based biomarkers for blinding eye disease

Computer vision and artificial intelligence provide unprecedented opportunities to realize new image-based biomarkers with the potential to drive precision drug development and personalized medicine; particularly in degenerative disease of eye and brain, where tissue sampling is not possible and where no treatments exists. This research project will deploy AI models for multiple tasks including segmentation, classification, anomaly identification and disease progression estimates taken from ophthalmic data collections across multiple clinics. Results of this project will provide the partner with critical human-machine interface to facilitate collaboration within and between clinical sites to enhance expert curation/annotation. The ability to identify unique patient subpopulations will allow improvements to drug development and clinical care.

Hardware and Software Integration for Automated Drone Surveillance

Unmanned Arial Surveillance is rapidly gaining acceptance for various applications, such as monitoring of long power transmission lines, pipelines and mass transit systems that extend for hundreds of kilometers. Unmanned Aerial Vehicles (UAVs) such as drones provide the flexibility to reduce costs. In the case of natural disaster occurrence such as earthquake, flood or hurricane, drones can quickly fly over to high risk areas where human access would be impossible or dangerous and provide information for rescue operations, etc. In this project, the goal is to develop an automated drone for surveillance purposes that can provide real-time monitoring for the target application. Hardware and software provided by AirMarket will be integrated in this project. The solution will be used to generate a framework for the next stage drone technology that can make drones fly beyond visual line of sight.