T3C: A traffic-communication coupling control approach for autonomous intersection management system

Abstract

Autonomous intersection management (AIM) system requires communication protocols with low delay and high reliability. However, most previous studies optimize the connected automated vehicle’s (CAV) communication and control systems individually, ignoring their collaboration and cascade effects. To address this gap, we present the Traffic-Communication Coupling Control (T3C) approach for joint optimization of CAV trajectories and communication networking. The roadside unit (RSU) periodic intervention mechanism and the edge-end collaborative computing architecture are utilized to adapt the AIM system’s multi-type computational tasks. The approach creates a relay CAV identity assignment module to provide a linkage pattern between communication networking and CAV control. Following that, CAVs utilize a distributed trajectory planning approach to plan their trajectory states, with parallel distributed model predictive control applied on a rolling horizon. The RSU collects and transmits the trajectory states to the mobile edge computing (MEC), which optimizes communication networking. To quickly solve the networking scheme, the task is divided into two sub-problems: backbone network generation based on the traffic-information flow coupling mechanism and information flow distribution. These two sub-problems are handled using the adjacency matrix masking optimization approach and enhanced adaptive large neighborhood search (ALNS) algorithm, respectively. Numerical studies are carried out to confirm the effectiveness of the proposed approach in various vehicle arrival rate scenarios. The results demonstrate that T3C can ensure stable low-delay communication while improving traffic efficiency, particularly in high vehicle arrival rate scenarios. Specifically, T3C achieves a low travel delay ratio of 28.38%–53.67% at the cost of an average transmission delay of 13.90 ms–24.95 ms.