MPC based path tracking control for autonomous vehicle with multi-constraints

In this paper, a nonlinear model predicted control algorithm is proposed for an autonomous vehicle with the known road information which is described by road curvature. The single track vehicle dynamic model is established and vehicle future state is predicted during the predict horizon. To minimize the tracking error, the standard quadratic cost function is proposed to calculate the optimal steering wheel and longitudinal force. Meanwhile, the anti-side slide and anti-rollover speed limit are considered into the constraint of the optimization. The simulation result illustrates that the proposed model predicted control has satisfactory path tracking performance and prevents the vehicle from unstable under different road conditions.