Safe Predictor-Feedback CACC with V2X-Aware Adaptive Spacing for Heterogeneous Vehicle Platoons

Abstract

Vehicle-to-everything (V2X)-enabled cooperative adaptive cruise control (CACC) is a key technology for improving both traffic efficiency and driving safety in vehicle-platooning scenarios. However, real-world platoons consist of heterogeneous vehicles with different actuation, computation, and mechanical delays, and communication latency also varies over time. Therefore, conventional approaches based on homogeneous vehicles and fixed-delay assumptions may fail to guarantee physical rear-end collision avoidance under severe driving conditions. This paper proposes Safe PF-CACC, a predictor-feedback-based CACC framework that integrates a V2X-aware safe inter-vehicle distance (Safe IV Distance) model with adaptive time-headway scheduling for heterogeneous vehicle platoons. The proposed Safe IV Distance is computed by considering communication latency, vehicle dynamic delays, and friction-dependent braking limits. It consists of three components: a minimum margin (MM) for low-speed and standstill conditions, a response-lag loss (RLL) induced by communication and vehicle dynamic delays, and a braking-performance limit (BPL) caused by road-friction-dependent braking capability. The resulting Safe IV Distance is converted into a dynamic effective time headway and incorporated into the predictor-feedback (PF) controller, while a filtering process is applied to suppress abrupt gain-scheduling variations. To evaluate the proposed framework, three representative CACC scenarios were considered: heterogeneous passenger-vehicle platooning, emergency vehicle platooning, and truck platooning. The simulation results show that overly short spacings without real-time delay awareness can cause collisions in high-speed and emergency driving scenarios, whereas overly conservative spacings improve safety at the cost of increased road occupancy. In the heterogeneous passenger-vehicle scenario, the proposed Safe PF-CACC reduces the maximum jerk and mean spacing by 20.6% and 49.4%, respectively, compared with the existing conservative method. In the emergency vehicle scenario, it achieved collision-free operation while reducing the maximum jerk and mean spacing by 18.6% and 53.2%, respectively. In the truck-platooning scenario, stable jerk and acceleration responses are maintained while the mean spacing is reduced by 59.6%. These results demonstrate that the proposed framework provides a practical integrated control approach for maintaining both control stability and physical safety in CACC systems under time-varying communication delays and road friction uncertainty.