Innovative Projects Realized

Visualizing meaningful changes in multivariate data of the hemodynamic response during cortical activation

This project involves working with researchers at the Bloorview Kids Rehab in Toronto to aid them in developing safer and more advanced brain]computer machines for individuals with severe physical limitations. The intern is developing a computer-based visual analytics tool that will help them further interpret the data they are collecting about brain activity. This tool will enable them to identify and compare change points in the data which could not previously be identified; a necessary step to developing advanced brain]computer machines.

Modelling User Web Browsing Behaviour for Relevant Content-based Retrieval

Building a mathematical model of how user behaves on a website is a useful method for determining what a user does and does not like. This model allows for information relevant to a user to be presented without having the user go out and find the information by themselves. These types of systems, generally known as “recommender systems” allow websites to provide custom content to users without a user having to explicitly state what they are interested in. WhoThaMan wants to develop such a system so that users are presented with sports news articles that are relevant to them. WhoThaMan then provides a forum for users to debate current events in the field of professional sports, thereby generating content on their site. Knowing what a user likes allows WhoThaMan to tailor the news articles to the individual user and increase the chances that a user will contribute their opinions to the site.

Managing Service Level Agreements in Distributed Systems

This project seeks to simplify the development and management of business processes deployed on a distributed Service Oriented Architecture (SOA). The target architecture is an enterprise system with distributed services coordinated by application workflows or business processes. Declarative goals, specified in Service Level Agreements (SLA), are used to assist in the development of such applications, and to automate the monitoring, deployment, and resource provisioning tasks. The complexity of developing and maintaining these processes is addressed by systematic SOA development cycles that identify the roles and specialities required by participants at each stage of the development cycle. While tools exist to assist in this process, the development, administration and maintenance of a business process still requires much manual effort that can be automated. In particular, the business process' nonfunctional goals, often expressed as Service Level Agreements (SLA), need to be manually considered at each stage of the development process. For example, completion time requirement for a process will influence decisions in the development, deployment and resource provisioning, and monitoring of the process. In our approach, the SLAs are captured and formalized early in the development process in a machine-understandable form. This allows the system to interpret, infer, automate various tasks. For example, an SLA on the required throughput of a process can automatically activate the instrumentation and generate the rules necessary to monitor the process throughput.

A Systematic Approach to Prioritize and Select Software Test Cases

Software testing in one of the imporant phases in software development process and has a significant impact on the final product quality. However, the spent cost and time for testing are high, and often increase after several releases. This project aims at test case selection and prioritization at Research in Motion (RIM) Company. The objective is to investigate the current testing process at TIM, and propose a systematic approach for improving efficacy and efficiency of test case selection and prioritization. For this purpose, available statistical and decision]theoretic methods will be investigated to find the most appropriate match to this problem domain at RIM. The ultimate outcome will not be necessarily automating these activities, and can be a recommending or decision]support product.

Physiological and psychological strain of firefighters during emergency response scenarios

Parallel Computational Weld Mechanics

Goldak Technologies Inc. (GTI), www.goldaktec.com, is an Ontario SME and worldwide leader in Computational Weld Mechanics (CWM), a critical technology for building nuclear reactors, ships, submarines, and offshore structures.  GTI’s software package VrWeld is arguably worldwide the best software for CWM.  The goal of this project is to build a new version of VrWeld, optimized for parallel processing, with at least 10-fold increase in computing speed.  This will enable novel industrial applications of VrWeld such as real-time control of industrial welding processes and improved design of experiments (DOE) for better accuracy.  The project will help expand GTI’s worldwide leadership position and market share in CWM, and will be of significant benefit to GTI Inc. and Ontario.

Characterization and amplification of blood cells derived from human dermal fibroblasts

Recently it was shown that cells taken from adult skin could be converted to cells possessing “stem-cell-like” characteristics by a process called cellular reprogramming.  However, the use of these cells is limited by our understanding of how to derive mature and functional cells, such as blood cells that could be used for disease treatment.  We previously showed the ability to generate mature blood cells directly from human skin fibroblasts without establishing a “stem cell-like” state, providing a safer option for cell-replacement therapies.  Our current project proposes to characterize the cellular-mechanism associated with mature blood cell formation from fibroblasts.  Furthermore, we aim to identify chemicals to increase blood cell formation and replace viral-delivery systems used to derive these cells from fibroblasts.  Therefore, the study in partnership with Ontario’s biotechnology/pharmaceutical sectors will provide an exciting new cell source that could be safely used for cell-transportation therapies to treat blood disorders.

Wavelet Methods for Optimal Control Problems in Pharmaceutical Research

Optimal control problems and reaction-diffusion systems have many applications in pharmaceutical research.  For instance, the enzymatics collagen matrix degradation and the principal process governing drug transport inside a solid tumor could both be described by reaction-diffusion systems.  Solving these systems using effective numerical algorithms will benefit drug product design, drug delivery and cancer treatment.  Since the traditional numerical solutions to these systems require large memory and long computational times, we will consider wavelet methods which allow us to combine high order accuracy, efficient preconditioning, and adaptive approximations.  We expect the wavelet methods to show the superb performance on solving the reaction-diffusion equations.  To further improve the numerical performance of the wavelet methods, we will implement multilevel techniques on adaptive grids.

Co-registered MRI and ultrasound imaging for improved detection and treatment of breast cancer

For this project, we will be improving a novel magnetic resonance (MRI) and ultrasound (US) imaging platform, which will improve diagnosis and treatment of breast cancer patients.  A specialized MRI-US tool has been developed by Sentinelle Medical Inc. (SMI) which helps guide US-imaging of tumours that are initially found using MRI.  The main goals of this project are to validate the accuracy of this tool and demonstrate its efficacy in the clinic.  MRI is commonly used for the detection of breast lesions, but non-cancerous conditions can have the appearance of cancer on MR images.  In these cases a biopsy is done, generally under US-guidance.  However, not all MRI-visible lesions are also visible on US.  A primary focus of this work will be to use SMI’s MRI-US tool to help identify these lesions that are invisible on US but can be seen on MRI.  The MRI-US tool will be used to help guide US biopsy of these lesions, and more advanced imaging techniques will also be examined for their potential to highlight tumours.  These will include techniques that highlight tumours based on their increased stiffness relative to the surrounding tissue and/or the highly vascular nature of cancerous tumours.  This project will improve our ability to care for cancer patients, using a novel multi-modality (MRI and US) system.  This work will directly benefit breast cancer patients, hopefully reducing the number of unnecessary biopsies and surgeries.  This will reduce stress on the individual patient, and it will also reduce overall cost to the health care system.

Advanced Design of Compact and Broadband Antennas for Vehicular Wireless Communications

Modern society has come to depend significantly on its wireless communication systems. However, the technologies for wireless communication still need to be improved further to satisfy the demands for higher capacity and higher data rates. In vehicular wireless communication systems, one of the key issues is the design of compact antennas while providing wideband performance. This project focuses on novel antenna design achieving very wide bandwidth or multi-band multi-functional performance, compact configuration, omni-directional radiation pattern and ease of fabrication. The novel antenna design will give the Canadian wireless industry and in particular my industrial partner (Sinclair Technologies Inc.) an edge in a market of formidable competition. Improved antenna performance means improved product quality and higher customer satisfaction. The engineers from Sinclair Technologies Inc. will interact with the McMaster team, which has world-class expertise in antenna design and optimization. They will be able to explore ideas related to novel antennas for next generation vehicle wireless communications. At the same time, I will have the unique opportunity to work on real-life high-complexity antenna design problems.

Optimization for Risk and Portfolio Management

The recent global financial crisis has made risk management a foremost concern of regulatory agencies and corporations worldwide. As the world's leading provider of enterprise risk solutions, Algorithmics Incorporated continues to showcase, and rely on, the talents of Ontario's R&D professionals in meeting the demands of the financial industry. Given the rapidly expanding scope and complexity of risk-aware management, mathematical innovation is central to Algorithmics' ongoing success. Notably, numerical optimization, which automates the construction of portfolios that best meet specified requirements, is finding novel uses in the field of risk management. The proposed research will study optimization problems that arise in two prominent applications. First, we will investigate ways to construct better replicating portfolios, which are used in the insurance industry as a proxy for liabilities to facilitate risk analysis. Second, we will examine optimization models for minimizing the credit risk of a portfolio of financial contracts.

Alternative Hydrogen Sources For Portable Fuel Cells

In the current climate of environmental awareness the need for alternative energy sources is undeniable.  In this respect hydrogen gas is a frontrunner as a clean burning fuel, with water being the only by-product.  Coupled with a PEM (polymer electrolyte membrane) fuel cell, hydrogen is a promising alternative to gasoline for automotive applications.  While significant progress has been made, this technology is still in its infancy.  A significant drawback in this technology lies in the storage and transportation of hydrogen within a vehicle.  To this end, our project focuses on the delivery of hydrogen to a fuel cell via the hydrolysis of chemical compounds bearing high hydrogen content.  Storage of hydrogen in molecular compounds is a safe and convenient alternative to storing hydrogen in heavy, high-pressure containers.  Moreover, in a typical internal combustion engine only ca. 30% of the energy from gasoline is utilized to propel a vehicle forward.  The remainder is lost as engine waste heat, which is removed by engine coolant.  An additional aspect of this project will be to utilize this wasted heat to power a thermally driven chemical reaction where hydrogen gas is released, which can then enter and power a fuel cell.  Recombination of the by-products from the chemical reaction within the fuel cell would regenerate the original hydrogen containing chemical compound and thereby complete a closed-loop in a thermally regenerative fuel cell.  Our initial target application for this technology will be for low efficiency transport trucks where significant amount of wasted heat is given off during long haul deliveries.