GameOpt: Optimal Real-time Multi-Agent Planning and Control for Dynamic Intersections


We propose GameOpt: a novel hybrid approach to cooperative intersection control for dynamic, multi-lane, unsignalized intersections. Unsignalized intersections are one of the more complex, prone-to-accident scenarios in modern transportation networks. Cooperation among Connected Autonomous Vehicles (CAVs) is a promising approach to unsignalized intersection control providing increased safety, efficiency and fairness. GameOpt is a hybrid formulation that first uses an auction mechanism to generate a priority entrance sequence for all the agents, followed by an optimization-based trajectory planner that computes the optimal velocity commands obeying the priority sequence. This coupling enables real-time performance in high density multi-agent traffic, simultaneously providing guarantees in terms of fairness, safety, and efficiency. Our approach can operate at real-time speeds (less than 10 milliseconds), which is at least 100 faster than other fully optimization-based methods. Tested on the SUMO simulator, our algorithm improves throughput by at least $5%, time taken to reach the goal by 75%, and fuel consumption by 33%, compared to auction-based approaches and signaled approaches using traffic-lights and stop signs.
Code Simulator Tech Report Proofs
Github(Coming Soon) SUMO Supplementary Report Proofs


Please cite our work if you found it useful,

      title={GAMEOPT: Optimal Real-time Multi-Agent Planning and Control at Dynamic Intersections}, 
      author={Nilesh Suriyarachchi and Rohan Chandra and John S. Baras and Dinesh Manocha},