Robust Autonomous Intersection Control Approach for Connected Autonomous Vehicles

Abstract

Traffic light-free intersection control is envisioned to alleviate congestion and manage vehicles intelligently. With the help of vehicle-to-infrastructure (V2I) communication and edge computing (EC), vehicles are instructed to cross the intersection with high vehicle safety and traffic efficiency without traffic lights. However, unstable channel conditions can lead to the reduction of traveling safety. In this paper, we propose a robust autonomous intersection control (AIC) approach with global optimization scheduling, which protects connected autonomous vehicles from collision under any channel conditions while achieving decent traffic efficiency. In particular, we propose an AIC model that gives vehicles certain autonomy under centralized control to ensure the traveling safety in case of some emergencies. By conducting an interference graph, we simplify the AIC problem as a weighted maximal clique problem with restriction. To improve the fairness and efficiency in terms of vehicle passage, multiple factors such as travel delay, traffic of the current lane and passengers' desired speed are considered. Furthermore, we propose a heuristic algorithm to search the solution space. For further optimization, a particle swarm optimization algorithm is proposed, achieving a near-optimal result with adjustable overhead. Finally, we build the simulation model and conduct a comparative performance evaluation. Simulation results demonstrate the superiority of our proposed scheme.

Keywords

• Connected autonomous vehicles
• autonomous intersection control
• edge computing
• robustness
• vehicular network
• optimization