Path-Tracking for Autonomous Vehicles with On-line Estimation of Axle Cornering Stiffnesses

To improve the accuracy and stability of path tracking under critical driving maneuvers, we propose a novel path tracking method for four-wheel-independent-drive electric vehicles (4WID-EV). An adaptive prediction model is applied to deal with the strong nonlinearity and time-varying behaviors of tires under critical driving maneuvers. The instantaneous cornering stifffiesses of the prediction model is estimated by an online estimator which combines unscented Kalman Filter (UKF) with recursive least square method with adaptive forgetting factor (adaptive FF-RLS). In addition, to handle the dynamics constraints under critical driving maneuvers, the model predictive control with multi-constraints (MMPC) is proposed to calculate the desired steering angle. The joint simulation results under two scenarios indicate that the proposed controller can minimize the path tracking deviations compared with the normal MPC controller, and maintain the stability of autonomous vehicle at high speeds.