Enhancing Driver-Automation Interaction Using RL-Based Shared Control

In autonomous driving, shared control requires optimal adjustment of the relative weight between human driver input and automation control to ensure safety and vehicle stability. This study proposes a Twin Delayed Deep Deterministic Policy Gradient (TD3) Reinforcement Learning (RL) based authority allocation approach incorporating driver behaviour to enhance adaptability and driver-automation collaboration. Simulations conducted in the MATLAB/SIMULINK-CarSim environment demonstrate that the proposed shared control framework significantly reduces lateral offset, heading error, and abrupt steering movements, leading to smoother control transitions and enhanced driving comfort. The results demonstrate the efficacy of the proposed method in enhancing the driver-automation interaction, ensuring a stable, intuitive, and safe shared driving experience.