Artificial Intelligence-Empowered Trajectory Tracking Control for Autonomous Vehicles: A Survey on Applications, Frameworks, and Future Trends

High-performance trajectory-tracking control is essential for high-level automated driving (AD), which demands superior tracking accuracy and robustness to dynamic environmental perturbations. Integrating artificial intelligence (AI) with classical control frameworks offers a potential solution to these challenging demands. This article systematically reviews the latest advances in AI-empowered trajectory tracking, providing a detailed analysis of these achievements from the perspective of control frameworks. We highlight three core AI-empowered modes: AI-enhanced dynamics learning, imitation-guided policy synthesis, and reinforcement-learning–augmented adaptive control. By fusing domain knowledge with data-driven plasticity, these methodologies address critical limitations associated with modeling, parameter optimization, and policy learning. Furthermore, this study evaluates practical deployment barriers, emphasizing simulation-to-reality transitions, safety constraints, and upstream uncertainty management. To advance the field, we outline a research roadmap to develop trajectory tracking from a passive, reactive execution task into a proactive cognitive control paradigm for complex open-world scenarios.