Evaluation of Decentralized Algorithms for Coordination of Autonomous Vehicles at Intersections

Connected Autonomous Vehicles (AVs) with Vehicle-to-Vehicle (V2V) communication are becoming an essential component of the transportation system. Self-driving cars have the potential to optimize the roads' traffic flow, fuel consumption and remove the possibility of human error and distractions. In these systems, all involved vehicles must be fully autonomous for maximum gain. However, a fully automated system requires major updates in the transportation and network infrastructure. In this paper, we investigate intelligent traffic control mechanisms for autonomous vehicles at intersections as a replacement of traditional intersection control (i.e traffic lights). Two well-cited decentralized optimization algorithms for cooperative vehicles are compared with realistic simulations in SUMO. We investigate the safety and feasibility of deploying the proposed algorithms in the real world. Further, we study the scalability and performance of the algorithms in the presence of communication impairments associated with wireless channels. This side-by-side comparison helps to gain insight into the strengths and limitations of these types of algorithms.