Cloud-Based Collision Avoidance Adaptive Cruise Control for Autonomous Vehicles Under External Disturbances With Token Bucket Shapers
Kaiqun Zhu, Zidong Wang, Derui Ding, Jun Hu, Hongli Dong
Abstract
This article addresses the real-time collision avoidance and adaptive cruise control (ACC) problems for autonomous vehicles within a cloud control platform. Collision avoidance and ACC are identified as essential components of autonomous driving as they directly impact the operational safety of the system. In the cloud control platform, a token bucket shaper is employed to regulate data transmission rates, ensuring the priority transmission of critical data while effectively preventing network congestion and cloud overload. The primary objective of this study is to achieve real-time collision avoidance ACC by comprehensively considering the influence of external noise disturbances and the token bucket shaper. Based on a signal smoothing method, a novel observer structure is first constructed to enhance state estimation performance. Then, an observer-based controller is designed to counteract process noise disturbances, thereby improving the control performance of the vehicle-following system. Subsequently, a real-time collision avoidance constraint index is formulated, and a method for solving this constraint is proposed, overcoming the effects of immeasurable states. The impact of the token bucket shaper and external noise disturbances on control performance is thoroughly analyzed, and sufficient conditions are derived to simultaneously ensure real-time collision avoidance and the bounded stability of vehicle-following error. Finally, the effectiveness of the proposed collision avoidance ACC algorithm is validated through a simulation example.