Abstract

This review paper explores the current state of AI-driven control systems in autonomous vehicles (AVs), focusing on key techniques such as reinforcement learning (RL), Proportional-Integral-Derivative (PID) control, and hybrid approaches that combine traditional and AI-driven methods. While these techniques have enabled significant AV technology advancements, safety, reliability, scalability, and real-time decision-making challenges persist. The paper proposes several future innovations, including advanced RL techniques, integration of machine learning with traditional control systems, Model Predictive Control (MPC) and AI fusion, enhanced sensor fusion, and human-AI collaboration. These innovations address existing limitations and enhance AV control systems' adaptability, decision-making, and overall performance. The review concludes by discussing the broader implications of these innovations for the future of autonomous vehicles. It offers recommendations for future research to advance the field further.