Innovative Projects Realized

Treatment

Edson designs and manufactures innovative machines used for case packaging of goods, food, pharmaceuticals, and other products into corrugated boxes. Machine steps include automatic feeding of the boxes, folding, adding the product, and closing the box top. Market pressures require client ability to swiftly change over to package a new product. Edson must also be able to rapid customize design and correctly manufacture for a potential new client machine sale. Towards this goal the project will create 3D solid part models and intelligent subassembly hierarchies. Mating points on parts will be identified geometric tolerancing analysis, and best assembly sequences will be documented. Modern Design for Manufactuability and Assembly (DFMA) methods, including Poka-Yoke, ergonomics, etc. will be emphasized. The success of the effort will be measured by internal video capture during actual assembly, time and error reduction, client feedback, etc.

Hybrid modeling framework for flood prediction

A hybrid computational framework for short-term flood prediction in urban watersheds (characterized by overland runoff) will be developed to improve prediction accuracy. The framework aims to accurately predict an event, e.g. flood or no-flood, as opposed to traditional methods which estimate water flow characteristics, e.g. 6 feet above flood stage. Successful early prediction of these events can help authorities to take appropriate mitigation measures and to minimize losses from the flooding. The framework requires only current and historical data on water levels and precipitation in the area of interest (such as those collected by TRCA’s flood monitoring gauge network). The development of the new framework will complement existing hydrologic simulation systems to improve and enhance services provided by TRCA to local agencies and public.

Speech and language technology development for next generation interactive voice services

Automated speech enabled human-computer interfaces have become essential for the competiveness of many Canadian industries. The demand for more natural, simplified, and low cost user interfaces has made voice enabling of telecommunications, automotive, and consumer electronics devices nearly ubiquitous. Commercially available automated intelligent assistants have been developed to guide customers through problem solving scenarios. The demand for high performance in these services has increased the pressure to improve the quality and reduce the costs associated with the underlying speech and language technology. However, research and development in the areas of automatic speech recognition, natural language processing, and dialog systems requires specialized expertise and significant linguistic resources. Nuance Communications is a world leader in providing speech based interactive services and has a substantial presence in Canada. The Telecommunications Group at McGill University is one of the top centers in North America for research in communications and speech and signal processing. These Mitacs internships will leverage the combined expertise of McGill and Nuance researchers to enable the Canadian driven development of a new generation of speech and language products and services.

Centre for Operations Excellence Summer Internship Cluster 2013

The COE Summer Internship Cluster consists of seven sub-projects sponsored by six different industry partners. Each sub-project represents an important challenge for its sponsor. These subprojects include understanding predictors of service variation and of return-to-work outcomes within the Physiotherapy program at WorkSafeBC; scheduling at Tree Island Industries; optimization of work flow for a set of activities for Telus; developing a technique for calculating and visualizing flight delay risk based on an analysis of data on past flight disruptions for AeroInfo; scheduling home care nurses and allied health professionals for Fraser Health; developing a simulation model of portering operations at a Fraser Health hospital to test the consequences on service delivery associated with centralizing or decentralizing; and analysing passenger movement through YVR airport in order to identify problem areas where passenger processing time can be reduced.

Modeling logistics for supply of bioenergy feedstock

The current version of the Integrated Biomass Supply Analysis (IBSAL) model is deterministic. As such, the uncertainty surrounding the results obtained cannot be studied or quantified. The focus of this research project is to incorporate uncertainty into the model in order to make it more realistic, provide additional insights into its properties and potentially suggest new types of solutions with corresponding margins of error and variability. The probabilistic version of the IBSAL model will allow users to test and explore more realistic sets of input assumptions and allow for the investigation of “what if” type of scenarios. The probabilistic version of the model will provide an opportunity to quantify spatial variability of economic, carbon and fuels costs in response to major driving variables such as crop residue yield and recovery rates.

GPS/Wi-Fi/ MEMS Sensors Integration for Indoor Navigation

GPS, MEMS sensors and Wi-Fi are three main candidate technologies for pedestrian navigation. GPS cannot provide reliable pedestrian navigation indoors because the GPS signals are degraded in indoor environments. MEMS sensors provide accurate but relative positioning solution for short time periods of standalone usages because of MEMS sensors’ errors characteristics. Wi-Fi can provide absolute location information by using pre-existing wireless infrastructures. However, Wi-Fi positioning requires special radio maps databases which is labour intensive and time consuming. Moreover, accurate orientation information from Wi-Fi cannot be provided in any of the existing methods. This research aims to improve the accuracy, reliability and accessibility of indoor navigation in two different ways for the current smart phones using GPS, Wi-Fi and MEMS sensors. Firstly, an algorithm is proposed to estimate the Wi-Fi heading by using clustering based and estimation based techniques. Secondly, when the GPS/MEMS sensors position solution is accurate, it is used to build the database for Wi-Fi positioning. Once the database is ready, Wi-Fi positioning is combined with the GPS/ MEMS sensors solution to provide the final robust positioning solution. The two algorithms will be used in the T-PN, the Trusted POSitioning Inc.'s Trusted Portable Navigator, for different real time systems such as Windows and Android based smart phones. The result will be a more robust and accurate navigation solution for deep indoor environments.

Residual white spruce growth responses following partial cutting treatments in boreal mixedwoods

The boreal mixedwood forests of the Abitibi region of Quebec are often characterized by a canopy of trembling aspen with shade tolerant species such as white spruce and balsam fir in the subcanopy. In many cases, harvesting has converted complex mixedwoods into simpler, even aged forests. Due to the differences in regeneration and tree growth, the white spruce is often of small merchantable size classes when the trembling aspen is ready for harvesting. Partial cutting is a harvesting method where only some canopy trees are removed while others are retained in various proportions as residual trees. This remaining forest cover can maintain vital ecosystem services and meet silvicultural objectives such as increased timber production. It is thought that by removing some overstorey competition via partial cutting, the residual trees will experience an increase in growth rate. The purpose of this project is to evaluate the growth responses of residual white spruce trees in mixedwood stands where four different proportions of trembling aspen were removed by partial cutting ten years ago. Few studies have evaluated longer term growth responses of residual trees in trembling aspen-white spruce mixedwood stands. It is essential to determine the effectiveness of partial cutting treatments in terms of growth response, since it will likely only be considered successful on an ecosystem and economic basis if the growth responses of residual trees are positive.

Information system for SCA

The goal of this project is to build an information system to help SCA(Saskatchewan Cattlemen’s Association) improve the information flow between itself and cattle producers, make information more accessible for cattle producers and automate some of the procedure SCA currently has to do manually. Examples include sale reporting via web services and cow tracking via mobile devices. In order to do so, the database of SCA needs to be redesigned and three mobile applications need to be developed. A more effective information system helps SCA to increase its productivity, reduce the chance of error and have better understanding of the beef market.

Web based avionics certification management for bus design

Aircraft/Avionics System certification is awarded by aviation regulatory bodies to applicants after they have demonstrated that the design of the system meets airworthiness requirements. Airworthiness requirements are a complex set of requirements, standards, and regulations. Demonstration of compliance to airworthiness requirements can be accomplished through testing, reviews, similarity of installation and/or usage and other means. Certification is, therefore, a complex and expensive process that involves designers, project managers, certification authorities, testers, aircraft, test benches, and other systems.

Marinvent is conducting a research program on a new Flight Management System (FMS) guidance algorithm to improve en route performance for its aircraft. Marinvent CertPro, a web based application to help managing certification programs will be used as a supporting tool for certification of the FMS en route algorithm.

Expected benefits for Marinvent are the ability to participate in cutting edge research programs once the FMS research program will be completed, and to propose CertPro as a certification management software product to the worldwide market, positioning Canada as the world leader for those technologies that have no competitor for the moment.

Predicting for Targeted Advertising by Maximum Information Coefficient

Predictive modeling is a statistical data mining approach that builds a prediction function from the observed data. The function is then used to estimate a value of a dependent variable for new data. Then objective of the project is to develop predictive models by machine learning approach and data mining techniques from the large volum of data collected at the partner site. The intension is to study applications of some most recent adademic reserach restuls to an industry product, a real-time analytics and data-intensive platfome. The resutls will be compared with the current system performance and have signaficent impact on the further development of the partner’s business.

Book Recommendation: Improving Collaborative Filtering with Content Information

Collaborative filtering is a product recommendation technique for making automated product suggestions to a user based on the preference information from similar users. Traditional recommendation algorithms drive personalized recommendations using the data from user purchases and ratings. For e-book retailers, besides user purchases and ratings, product features such as book content and metadata also provide valuable information that can be used 'to improve the recommendation precision and recall. The goal of this project is to improve the recommendation algorithm in the e-book domain by incorporating additional user. and product features. Research topics include the application of collaborative filtering, topic modeling and collaborative topic regression as well as experimenting with novel approaches.

Occurrence of the Timiskaming-type sedimentary rocks in the North Caribou Greenstone belt

Jason Duff, who is completing MSc program at the University of Ottawa, recently found evidence for sedimentary rocks in the North Caribou greenstone belt that are 2680 million years old, younger than other rocks in the belt. The rocks elsewhere in the belt contain 3.0 billion year sedimentary and volcanic rocks including the host rocks for the Musselwhite mine. Similar young sedimentary rocks, called Timiskaming-type rocks, are known to be spatially associated with large gold deposits elsewhere in the Archean Superior Province (which forms a large part of “Canadian Shield”), including the Porcupine, Kirkland Lake, Larder Lake, Malartic, Pickle Lake, and Hemlo gold camps. The proposed study will determine the distribution of this young sedimentary rock unit in the belt and evaluate the source of the sedimentary rocks; outcrops in the area will be examined, and rock types identified and their ages determined by isotopic methods. This finding of relatively young and immature sedimentary rocks in the belt will help guide exploration for gold deposits in this area.