Hybrid Model-Based and Data-Driven Sliding Mode Control for Robust Path Tracking in Intelligent Vehicles

Robust path tracking for intelligent vehicles represents a critical challenge due to model uncertainties and varying operational conditions. This paper proposes a novel model-data-driven sliding mode control (MDD-SMC) framework to address this issue. The core of this framework is a hybrid controller architecture. A model-based law, founded on SMC principles, governs the nominal vehicle dynamics to ensure baseline stability and robustness. This is synergistically integrated with a data-driven compensator, designed based on model-free adaptive control theory, which actively learns and rejects composite disturbances arising from unmodeled dynamics, parameter errors, and external interferences. Furthermore, a guidance law is employed to fuse lateral and heading errors into a single composite variable, simplifying the control problem to a provably stable single-input single-output system. Comparative simulations demonstrate that the proposed MDD-SMC consistently achieves superior tracking accuracy and stability over conventional control methods, particularly in challenging high-speed and low-adhesion scenarios. This framework therefore provides a highly adaptive and effective solution, significantly enhancing vehicle safety and control precision.