Development and Evaluation of an Adaptive Traffic Control System
Development and Evaluation of an Adaptive Traffic Control System for Urban Networks Based on Video Sensor Data
This project is to address two technical challenges faced by the industrial partner in the process of developing a new video-based, adaptive intersection traffic signal control system.
The first challenge is to determine the current state of the traffic network, such as link volumes, turning movements, size and speed of platoons, based on the company’s video sensor data. The video traffic data includes traffic volumes entering and exiting each approach of an intersection being monitored, which by itself is incomplete for the purpose of traffic signal control. The missing state variables along with those of the intersections without any camera must be estimated using models so that all traffic in the controlled area are considered in signal optimization.
The second challenge is how to predict the future state of the traffic network based on the current traffic state, historical traffic patterns, and a proposed signal control plan. Without a reliable prediction model, the signal optimization process could just yield suboptimal myopic solutions. This Engage project will help the company develop a truly innovative traffic management solution, which will enhance the company’s position to become a leading traffic solution provider in Canada. The new traffic control system is expected to benefit all Canadians by minimizing traffic delay, improving intersection safety, and reducing pollution and fuel consumptions.