The distinguished lecture – AI-Driven Connected Transportation: Moving Beyond Efficiency Towards Sustainability by Prof. Shen Wang, at 12.00 pm on 25th October 2024. The topic for the lecture on the cooperative control for traffic light signals and connected autonomous vehicles using deep reinforcement learning. In order to be consistent with the United Nations Sustainable Development Goals (UN-SDGs) and support the development of future smart transportation systems, in addition to reducing travel time, further reducing fuel and emissions, improving traffic safety, and simplifying infrastructure deployment and maintenance should also be considered. Professor Wang introduced a multi-agent Deep Reinforcement Learning (DRL) system. The system collaboratively controls traffic lights and Connected Autonomous Vehicles (CAVs) to well balance the reduction of travel time, fuel, and emissions. The system can also be extended to complex urban scenarios by cooperating with only one CAV nearest to the traffic light controller on each incoming road. This avoids expensive coordination between the traffic light controller and all possible CAVs, leading to stable convergence of training CoTV in large-scale multi-agent scenarios.

At the beginning of this event, Zixuan Jia, Chair of the University of Birmingham IEEE Student Branch, gave a brief overview of the Student Branch and introduced the activities held in the past and the functions of Student Branch officers.

And then, the host, Prof. Weiqi Hua kicked off this event, welcomed everyone and introduced the speaker, Prof. Shen Wang to the audience. During the presentation, Prof. Shen Wang introduced the multi-agent deep reinforcement learning system called CoTV with the ability of cooperative controling the traffic lights and the speed of vehicles. Also, the explainable robust algorithm for privacy-preserved federated learning in future networks to defend against poisoning attacks called SHERPA was presented. Futhermore, Prof. Shen Wang shared the audiences with their on-going research, including deep reinforcement learning for motorway traffic, LLM-based explainable AI, and dynamic princing for electric vehicle charging.

The participants engaged actively in the questions and discussions part with unique thoughts and innovative perspectives to the lecture. Prof. Wang responded to their questions with patience and openly shared his insights.

By Yixin Li,

Vice-Chair, UoB IEEE Student Branch