Appropriate sensor signal analysis and abstraction in physical activity game design

The industrial partner, Digido, is interested in developing and marketing exercise games targeted at children that leverage the increased prevalence of  smartphones and their sensing and computational capacity, including: their ability to detect activity levels; their increasing use as a gaming platform; and their integration with social media and online communities. However, current activity sensing and classification techniques are too limiting for use in smartphone-based exergames for children.

Maneuvering Simulation of Planing Hull

With dramatic improvements in vessel performance and tactical systems of high speed crafts in recent years, naval, coast guard and law enforcement agencies increasingly task them to complete a growing range of operational objectives. The combination of faster vessels, more sophisticated systems and extended responsibilities has driven fleet operators to re-examine how their boat crews are trained.

Novel 3-D User Interfaces for improved situation awareness and mobile robot control

In an alien or possibly hostile environment, the situation awareness of a remote robot operator will be limited. Map information may not be known beforehand. The site may also be in a dynamic state where changes occur in the surrounding in any moment. The main objective of this project is to develop novel technologies to increase situation awareness of remote robot operators and their ability to intuitively interact with the robots for more efficient operations.

Distributed collaborative recommendation engine for Asset Store


In this project we attempt to research and develop from ground up a scalable distributed computing based recommendation engine using machine learning. A computer science student from the University of Toronto will work with Side Effects Software at their Toronto office to implement the research intensive recommendation engine algorithm and integrate it in the smart asset online store. We expect and hope that this will result in high quality recommendation, is scalable and has a strong foundation in statistical machine learning based algorithm approach.

Online Risk-Driven Management Framework for Territorial Security in Wireless Sensor and Robot Networks

Small teams of mobile robots provide nowadays the ability to assist wireless sensor networks in many threatening scenarios that unexpectedly arise during their operational lifetime. The perceived risk or vulnerability that the network is exposed to triggers an immediate, corporate action from the robotic agents (actuators). We focus on a sort of robots which are able to carry static sensors and deploy them all over the field.

Using LSA for Automatically Assessing Free Texts

To evaluate the content of free texts is a challenging task for human. Latent Semantic Analysis (LSA) can be used to automate this process. The main idea behind the LSA technology is to extract the close relationship between the meaning of a text and the words that are present in that text. ShirWin Knowledge and Learning Systems Inc.

Generation of Application Level Traffic Markings For Smartphones

Mobile Network Carriers are experiencing unprecedented data traffic loads which are straining their existing infrastructure. They are thus interested in exploring research areas which focus on reducing this traffic load. The malicious traffic generated by malware on smartphones is of particular interest. My research focuses on the generation of a software system that marks traffic from smartphones to indicate the specific phone application which has generated the data flow.

Precursor charge prediction for improved peptide identification with mass spectrometry

The research project aims to develop an effective method that utilizes multiple features to improve mass spectrometry based peptide identification with database search approach. The project is a continuation to the student’s previous research on precursor charge state prediction, since predicted charge state is a novel feature and has a great potential to discriminate the correct and incorrect peptide identifications.