Yazdani Bejarbaneh, Elham (University of Wollongong), Du, Haiping (University of Wollongong), Naghdy, Fazel (University of Wollongong)

Data-Driven Optimal Cooperative Control for Vehicle Platoons with Unknown Dynamics

Scheduled for presentation during the Invited Session "Enhancing Trustworthiness and Resilience of Connected and Autonomous Vehicles in Adversarial Environments" (FrAT7), Friday, September 27, 2024, 10:30−10:50, Salon 15

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on October 8, 2024

Abstract

Platooning control of connected and automated vehicles (CAVs) has emerged as a promising strategy to enhance fuel economy, traffic efficiency and safety. One of the main challenges in coordinating multiple vehicles is executing safe and efficient cooperative control, while considering the inherent nonlinearities in vehicle dynamics. This study aims to design a data-driven optimal distributed leader-tracking control for heterogeneous vehicle platoon in the traffic scenario. Every vehicle within the platoon is treated as a multi-input-multi-output system, exhibiting nonlinear dynamics. The distributed control system is developed using reinforcement learning (RL) algorithm that learns the optimal control policies from the vehicle system data without requiring prior knowledge of vehicle dynamics. This method ensures that all platoon vehicles reliably follow the leader’s behaviour, even in the presence of unknown and nonlinear dynamics. Simulation results involving a platoon of heterogeneous and nonlinear CAVs demonstrate the effectiveness of the proposed cooperative controller.