Model Predictive Trajectory Tracking Based on Extended Kalman Filter

Vehicle trajectory tracking is one of the core technologies in the field of automatic driving, various control algorithms are currently used in trajectory tracking. Among existing algorithms, model predictive control (MPC) outperforms other algorithms because it consists of model prediction, receding optimization, and feedback correction. Previous studies have designed trajectory-tracking algorithms under stable road conditions. However, the actual driving scenario is no longer stable, meaning those algorithms may not be applicable. Therefore, we establish a kinematic model for the vehicle with additive noises and propose a trajectory-tracking algorithm based on model predictive control and the extended Kalman filter (EKF). The algorithm can still guarantee real-time calculation and high tracking accuracy under unstable road conditions.