Cascade LPV Control for Automated Vehicle Trajectory Tracking Considering Parametric Uncertainty and Varying Velocity

Abstract

Trajectory tracking control is very crucial for autonomous vehicles (AVs). However, its performance can be degraded due to the time-varying velocities of AVs and their parametric uncertainties. To provide an accurate and smooth trajectory tracking effect under different driving conditions with varying velocities, a cascade Linear Parameter Varying (LPV) vehicle integrated control method by considering environmental uncertainties is proposed. Firstly, both kinematic and dynamic models are established using the polytopic uncertainty method with finite vertices to represent the variations of vehicle dynamics and the uncertain tire stiffness. The selected variables in each model are defined as the scheduling variables to describe the non-linearity of the vehicle model. Then, the LPV-based Model Predictive Control (MPC) and a Linear Matrix Inequality (LMI)-based Linear Quadratic Regulator (LQR) are designed to track the desired path in terms of kinematic variables and dynamic variables, respectively. Finally, the simulation results demonstrate that the proposed cascade LPV integrated control can accurately adn effectively track the planned trajectory.