A comparative study on trajectory tracking control methods for automated vehicles

Abstract

Abstract Trajectory tracking control, as the core component of the autopilot system, is essential for achieving high-performance autopilot functionality. Therefore, to design a trajectory-tracking controller that is stable, reliable, accurate, fast, and robust, this paper proposes three lateral control schemes and two longitudinal control schemes for comparative analysis. In lateral control, a lateral control strategy of linear quadratic regulator (LQR) based on feedforward and feedback is designed. Subsequently, a lateral controller utilizing a linear parametric time-varying model predictive control (MPC) approach is proposed, and the constrained optimization problem is addressed using a quadratic programming (QP) solver. Finally, a lateral control method based on nonlinear integral sliding mode control (NISMC) is developed, which realizes zero steady-state error by substituting saturation function for symbolic function and introducing integral action. The MPC and PID speed-tracking controllers are compared within the longitudinal control scheme. The effectiveness of these controllers and their respective performance metrics are validated through step speed tracking tests. Furthermore, the efficacy of the three proposed controllers, along with their advantages, disadvantages, and applications, is assessed by designing three distinct road surface adhesion coefficients for the double lane change test, and the circular path test under medium-low adhesion conditions. At the same time, in order to display the control characteristics of the three controllers more directly, this paper designed a 5 score chart according to various performance indicators listed in Table 2 to quantitatively compare and analyze the comprehensive performance of various controllers.

Keywords

• Trajectory tracking control
• Autonomous vehicle
• Linear quadratic regulator
• Model predictive control
• Nonlinear integral sliding mode control