Machine Learning for the Telecommunication Industry

Ericsson is an industry leader in offering telecommunication solutions and products. As an important step on the path towards the automatic and autonomous management of next generation networks, Ericsson is developing technology in machine learning and artificial intelligence that will benefit operators around the world, including in Canada where Ericsson supplies technology to most of the major telecommunication network operators.

Computational Lens-free Holography for Rapid Monitoring and Characterization of Airborne Particles

Air pollution is a major environmental risk to human health, and air quality has become an increasing concern in the industrialized world. Rapid and accurate detection and characterization of airborne particulates is crucial for monitoring and improving air quality. In this proposal, we develop a compact, cost-effective, computational lens-free holography platform for high-throughput characterization of airborne particulates.

Proof of Principle Application for Biosensors in Animal Models

NXTSens has developed an implantable continuous monitoring, stand-alone and potentially mass-producible microsensor for accurate monitoring of tissue pressure and temperature in damaged limb muscles to better diagnose ACS. The proposed research will involve several sub-projects that include: i) develop appropriate surgical models for the accurate and reproducible placement of the microsensors in rodent compartments; ii) test the performance of a prototype microsensor in a relevant cadaver model.

Evaluation of Alternative Battery Technologies for Adoption in Manitoba Hydro Substations

This project will investigate the applicability and merits of battery chemistries for use in a utility substation, where longevity, reliability, and security are prime considerations. The project aims to characterize the process of battery aging when batteries are used under representative utility substation loading profiles in order to determine how fast the batteries will age and what signatures may be used to determine how close the batteries are to the end of their life so that pro-active maintenance work may be initiated.

Implementation, Demonstration, and Evaluation of a Cloud-Based Smart Dual Fuel Switching System (SDFSS) for the Hybrid Integrated HVAC System in a Net-Zero Energy House

Energy consumption of a net zero energy home (NZEH) equipped with high efficiency natural gas fired furnace and an electrical air source heat pump will be monitored for a period of 12 months. The mechanical system of the house is designed to switch between these two sources of energy based on outdoor temperature. Based on current settings, heat pump works during milder/warmer weather condition and when temperature reaches below a certain point, system switches to furnace which works more efficiently in cold weather.

Topology validation, Error detection, and Correction of rooftops 3D models from LiDAR point clouds and Photogrammetry

The last few years have seen a tremendous increase of the collection and use of LiDAR (Light Detection and Ranging) data for 3D modeling of cities, forest surveys and autonomous driving. In the context of 3D cities, buildings are reconstructed separately, sometimes with the help of aerial images. The accuracy assessment of these 3D models is not straightforward. However, end-point users need to know how reliable the models are for using them for tasks such as roofing quotes and solar installations design.

Ransomware Detection through Device and Network Behavior Monitoring

Ransomware consists of malicious software that after infecting a target device prevents the device owner from using effectively the corresponding resources until the demands of the ransomware operator are met usually by paying a ransom, typically using cryptocurrencies.
Despite the growing number of ransomware infections, their increasing sophistication, and their significant financial and operational impact, available defensive mechanisms are still embryonic.

Development of a Co-Simulation Platform for Electrical Systems

The proposed research aims to develop better computer simulation tools for the study of large electrical power systems. The premise of the research is based upon the concept of co-simulation, wherein two specialized computer simulation tools, each with unique features and strengths, are used in conjunction to solve a large electrical system. In the particular case of the proposed research work, this will be achieved using an interface between an electromagnetic transient (EMT) simulator and a dynamic-phasor (DP) simulator.

Dynamic Binary Instrumentation of Embedded Systems

The proposed research project focuses on the dynamic analysis of embedded systems. The project has a duration of six months and aims to achieve two objectives: (1) to evaluate the applicability and effectiveness of a newly proposed QoS-aware dynamic instrumentation framework on real-word time-sensitive applications, and (2) compare the outcome to that of an existing state-of-the-art framework. For this purpose, the intern will get access to real-world applications provided by the partner, Labforge, Inc.

Quality Assessment and Enhancement of Retinal Images

Babies who are born prematurely are at risk of developing a condition called Retinopathy of Prematurity (RoP), which if left untreated, can lead to permanent blindness. RoP causes characteristic changes in the retinal vasculature,
which can be seen when looking into the eye. Because the infants need to be monitored regularly for this condition, and certain traits need to be carefully identified, a special camera is used to take a picture of the retina. These pictures can then be studied for signs of RoP by an ophthalmologist.