Implementing Trajectory Tracking Control Algorithm for Autonomous Vehicles

This paper adopts the lateral and longitudinal control decoupling strategy to study the control algorithm of automated driving systems. The Frenet-frame is used for lateral control. According to the lateral error and heading error of the self-driving vehicle and the reference path, the model predictive control (MPC) is used to realize the lateral control of the self-driving vehicle to complete the tracking of the target trajectory. Longitudinal control is separated into two modes: speed control and tracking control. According to the driving environment of autonomous vehicles, a decision-making control strategy is designed to achieve smooth switching of the controller for the two control modes. The speed control mode uses the proportional-integral-derivative (PID) controller to realize the speed tracking control of the reference speed or the driver's setting. The following error model is established for the tracking control mode, and the Linear-Quadratic-Regulator (LQR) controller is used to realize the distance control of the vehicle in front. Finally, the algorithm is validated based on the ROS-CoppeliaSim simulation platform and field testing of autonomous vehicles in the actual tournament. The results show that the lateral and longitudinal decoupling algorithm has a good control effect, environmental adaptability, and stability.