Preview control algorithm has been widely implemented in intelligent vehicle path-tracking controllers. The key challenge of developing such control is to determine the appropriate preview distance, which plays a vital role in achieving the optimal trade-off between two competing control objectives, tracking accuracy and driving stability. Additionally, vehicle speed and road radius have a significant impact on the optimal preview distance. Thus a hierarchical vehicle path-tracking control strategy based on the adaptive two-point preview is proposed in this article. In the upper-layer module, the two-point preview driver model is constructed to obtain the target yaw rate according to the comprehensive deviation. In the lower-layer module, the neural network sliding mode controller is employed to track the yaw rate and, therefore, achieve intelligent vehicle self-tracking. Moreover, the hybrid particle swarm optimization (PSO), combining the traditional PSO and genetic algorithm (GA), is proposed to optimize the preview distance with consideration of the tracking accuracy and driving stability. The resulting optimal preview distance is tabulated with inputs of the vehicle speed and road radius and utilized online to realize the adaptive change of preview distance. Finally, the co-simulation platform, coupling Simulink with Carsim, is employed to compare the proposed optimal hierarchical path-tracking control with the traditional control of which the preview distance is a fixed function of vehicle speed. The simulation results show that the adaptiveness of preview distance in the proposed algorithm significantly improves tracking accuracy. Compared to a traditional preview control of which the preview distance is a fixed function of vehicle speed, the maximum lateral deviation decreases 7.2%, 17.9%, and 22.3% at a vehicle speed of 20 km/h, 40 km/h, and 50 km/h, respectively.
