You are here

Fusion and Inference in Surveillance Networks

Project Type: 
Past

With the widespread deployment of networked sensors and cameras throughout cities, there is an incredible opportunity for improving safety and security. Surveillance networks incorporate cameras mounted on traffic lights and overpasses, mobile cameras attached to emergency vehicles, and chemical and biological sensors for detecting dangerous contaminants.

Project Leader(s): 

Dr. Mark Coates, McGill University

With the widespread deployment of networked sensors and cameras throughout cities, there is an incredible opportunity for improving safety and security. Surveillance networks incorporate cameras mounted on traffic lights and overpasses, mobile cameras attached to emergency vehicles, and chemical and biological sensors for detecting dangerous contaminants. Surveillance networks can comprise several thousand sensors and cameras throughout a city. The task of controlling and utilizing the information provided by these networks is a challenging engineering task as the mathematics and algorithms needed do not exist. Thus, the design of strategies for managing the sensors and cameras in the network and harnessing the information provided by them is the goal of this project. In the past year, the team devised a novel method for tracking multiple vehicles, as well as a new algorithm that allows networks to detect traffic jams and provide an early indication of accidents and their impact on traffic flow. A new approach for organizing automatic calibration of sensor networks and mechanisms to provide efficient synchronization of distributed computer networks were described.

Project team: 
Dr. Nando de Freitas, University of British Columbia
Dr. Arnaud Doucet, University of British Columbia
Dr. Frank Ferrie, McGill University
Dr. T. Kirubarajan, McMaster University
Dr. Michael Rabbat, McGill University
Non-academic participants: 
Funding period: 
October 1, 2021 - March 31, 2021