Abstract: Path tracking is a crucial technology in the field of autonomous driving; however, communication delays between sensors and actuators, as well as steering system response lag, can significantly impair tracking accuracy. To address this issue, an equivalent first-order linear model of the steering delay is constructed based on an in-depth analysis of its impact on path tracking precision. A feedforward-feedback integrated controller is designed, incorporating previewed curvature of the target path and compensation for the steering delay, thereby enabling high-precision path tracking under various speeds and delay conditions. A co-simulation model of the autonomous vehicle integrating steering delay characteristics is developed using Matlab/Simulink and Carsim. Comparative analysis of path tracking performance under different control algorithms demonstrates that the proposed integrated controller achieves significant improvements: at low speed, the performance metric is improved by 97.50%, 82.27%, 60.71%, and 95.56% compared to the traditional LQR controller, preview feedforward controller, delay-compensated feedback controller, and sliding mode controller, respectively; at medium speed, the improvements are 94.38%, 72.31%, 48.12%, and 93.33%; and at high speed, the improvements are 77.51%, 72.13%, 75.06%, and 88.04%, indicating superior path tracking performance.

Key words: steering delay, path tracking, preview-feedforward control, optimal control method, feedback control