Intelligent vehicle platoon driving control method and system based on security reinforcement learning

CN122290371BActive Publication Date: 2026-07-21CHANGAN UNIV

View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office: CN · China
Patent Type: Patents(China)
Current Assignee / Owner: CHANGAN UNIV
Filing Date: 2026-05-27
Publication Date: 2026-07-21

AI Technical Summary

Technical Problem

Existing platoon control methods struggle to balance reducing spacing and ensuring safety when dealing with highly dynamic and tightly coupled vehicle platooning systems. Factors such as sensor noise and sudden speed changes in the lead vehicle affect the lateral and longitudinal stability of the platoon, which can easily lead to traffic accidents.

Method used

A safety-based reinforcement learning-based intelligent vehicle platooning control method is adopted. The vehicle state is obtained through the communication topology of the leading vehicle, the navigating vehicle, and the following vehicles. A constrained Markov decision process is constructed, and the collision risk is separated into a cost function. The policy parameters are optimized by using the Actor network and the Critic network to achieve a balance between lateral and longitudinal stability and driving safety.

Benefits of technology

While ensuring convoy stability, it improves following efficiency and driving safety, and reduces collision risk through variable vehicle spacing strategy and safety reinforcement learning model optimization.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure CN122290371B_ABST

Patent Text Reader

Abstract

The application discloses a safe reinforcement learning-based intelligent vehicle platoon driving control method and system, relates to the technical field of automatic driving, and comprises the following steps: obtaining the actual driving state of platoon vehicles and the surrounding traffic environment state based on a preceding vehicle-leader vehicle-follower vehicle communication topology structure, setting the car-following strategy of a follower vehicle to determine the corresponding desired distance, calculating the trajectory tracking error according to the actual driving state and the desired distance, constructing an observation state space based on the surrounding traffic environment state, the actual driving state and the trajectory tracking error, modeling the platoon vehicle driving control process as a constrained Markov decision process, respectively constructing a reward function and a cost function, inputting the observation state space at the current moment into a safe reinforcement learning model, and outputting an action instruction; the method has the beneficial effects that the platoon driving control strategy can guarantee the stability of the lateral and longitudinal strings, and simultaneously consider the following quality and driving safety.

Need to check novelty before this filing date? Find Prior Art