An Open Learning Environment for Smart Transportation

With the recent advances in artificial intelligence, applying deep reinforcement learning to improve urban traffic efficiency and reduce traffic congestion has been gaining increasing interest in both academia and industry. This research program aims at developing computational platforms to evaluate models and algorithms for the next generation traffic control and management strategies, such as autonomous vehicles, vehicle-to-vehicle communication, and vehicle-to-infrastructure communication. The McGill team and Fundway Technology will develop a reinforcement learning platform based on real-world traffic data and scenarios. The other objective of this program is to develop a full-scale experiment based on Xuancheng city in China and provide researchers and practitioners an opportunity to implement their algorithms in a real-world feedback system.

Yinan Wang
Xudong Wang
Faculty Supervisor: 
Lijun Sun
Partner University: