Waypoint tracking predictive control with vehicle speed variation

Abstract

This paper presents a waypoint tracking method using model predictive control (MPC) for an autonomous driving vehicle. To solve vehicle speed variation in the MPC design, we propose a scheme using horizon-wise constant (HWC) MPC. The HWC MPC assumes that a prediction model uses a fixed parameter prediction model within a prediction horizon to resolve a parameter variation problem. To evaluate the performance of the proposed method, we performed three comparative studies with one speed variation MPC and two constant speed MPCs on a highway curve road. We verify the utility of the proposed MPC using HWC modeling and MPC using fixed velocity modeling. We have also confirmed that the proposed HWC MPC outperforms MPC at with fixed speed. The proposed controller is expected to be a solution to various parameter situations in waypoint tracking for the autonomous driving vehicle.