Optimal Model Predictive Control for Path Tracking of Autonomous Vehicle

Abstract

This paper investigates the optimal model predictive control for the path tracking of an autonomous vehicle. For the control algorithm of an autonomous vehicle following pre-calculated path, two indices must be considered: 1. The performance index including the path tracking error and the energy consumed in control process. 2. The computational cost. In this paper, an optimal model predictive controller is established by searching dynamically optimal linear model parameters when the model mismatch happens. Compared with the general nonlinear model predictive controller or the approximate linear model predictive controller, it is more suitable for path tracking of the autonomous vehicle due to the obvious reduction of computational burden and the good path tracking performance.