Development of a system for the automatic recognition and classification of normal and abnormal cells in human blood samples

Automation of medical diagnosis/detection process is very important in terms of enhancing diagnostic accuracy, increasing throughput, reducing cost, and training new staff. Our current goal is to go from the proof of concept stage (automatic recognition and classification of human blood images) to a complete working optimized prototype and to start testing it in an actual clinical lab environment with help from CLS. The prototype design will take into account user friendliness, high throughput, robustness, integration with existing lab work flow and reasonable cost.

Anti-islanding detection for renewable energy systems in distribution system

Integration of renewable energy systems into grid is an effective solution to the electric energy shortage and environmental pollution. A number of technical challenges may arise with increased grid-connected renewable energy systems. One of the most important issues is how to achieve the islanding protection. Many anti-islanding detection methods have been reported for single renewable energy systems in the last decades. In practice, however, the multi-unit systems are distributed in different feeders. Consequently, all of the existing methods might fail in this case.

Virtualized Wireless Access in the Green Sustainable Telcommunication Cloud System

This project is primarily focused on virtualizing wireless access network so that multiple operators can share the same physical resources while being able to stay isolated from each other. The basic idea is to allow wireless access points from different operators form a single virtual access point that efficiently manages its available resources. The goal is to exploit the advantages that can be obtained from virtualizing the air interface (i.e., spectrum sharing), protocol virtualization and flow-based virtualization.

Development of spectroscopic imaging technology for grain quality inspection

Fusarium fungi infestation causes Canadian grain producers a loss of almost $1 billion dollars per year. Fusarium fungi produce toxins, e.g., deoxynivalenol (DON) which cause toxic effects in animals and possibly humans. We will develop a portable hand-held hyperspectral imaging device to detect, in the field, Fusarium infestation in grains. We will also evaluate the applicability of spectroscopic Optical Coherence Tomography to accurately and quickly determine DON level in grains with high sensitivity (1 ppm to 10 ppm).

HVDC “Superline” for improved angular stabilityof AC-DC system

High Voltage DC Transmission (HVDC) is used for bulk power transfer over long distances. Manitoba Hydro system involves collection of AC power in the north where it is converted to DC and then transferred to southern Manitoba (approx 900km) where it is converted back to AC to feed consumers. Recently the problems of inter area oscillations have been reported in the system. Manitoba Hydro uses feedback signals such as frequency from the converter stations as control input to modulate power through HVDC links to damp these oscillations.

Calibration, Characterization and Optimization of Microwave Imaging System for Grain Monitoring

Microwave Imaging (MWI) is an emerging modality where the goal is to estimate the electrical properties of an object-of-interest. This is done by transmitting a microwave signal into the OI and collecting measurements outside the OI. The measurements are inputs to an optimization algorithm that solves for the unknown electrical properties. It has been proven using computational techniques that the proposed modality can be successfully adapted for monitoring moisture content inside grain bins.

Choosing the game and the stakes in a casino: A service pricing problem with many substitutable services

At any given time, a casino has to decide which table games to present to its customers and what betting limits to place on those tables. Understanding that a table games is a service and that the table limit is pricing, this is a classic service pricing problem. However, the number of tables and the fact that we can assign many different games to them leads to the novel research question: How do you which services to offer at what prices when customers have different preferences for each option?

Understanding Event-based Interactions in Enterprise Web Applications

JavaScript is widely used today to create interactive web applications that replace many traditional desktop applications. However, understanding the behaviour of web applications is a challenging endeavour for developers. Despite its importance and challenges, there is currently not much research dedicated to supporting program comprehension for web applications and popular industrial tools, such as Firebug and Chrome DevTools, are limited in their capabilities to support web developers effectively.

Wireless Asset Tracking and Locating System

Proposed research project consists of finding a suitable solution to the problem of tracking assets in environments with different constraints. The solution is expected to include the development of geolocation algorithm and then based on the algorithm creating schematics of several modules with transceivers, embedding of required low level software to the selected microcontroller controlling the transceivers, developing high level software for server side data management and client side monitoring and control.

Low-Power Signal Processing Algorithms for an Athletic Performance Monitor

GestureLogic is building a product that optimizes athletic performance. The product is a wearable sensory network that monitors muscle activity. The goal of this research is to take this inherently complex muscle data that is acquired by the sensory network and translate it to useful biometrics for the consumer with the help of intelligent algorithms. The algorithms will help intuitively visualize important metrics such as strength of muscle contractions, heart rate and fatigue. The benefit to the partner organization is twofold.