Model predictive control based on Kautz function for autonomous ground vehicle path following control application

This paper introduces and presents model predictive control approach based on Kautz function algorithm for path following control in autonomous car vehicle system. Predictive control is designed based on linearized vehicle and tire model, while the real vehicle is included a roll dynamic motion. The controller based on Kautz functions is compared with conventional method which is called Delta function in order to follow the desired trajectory while rejecting the crosswind. The vehicle is performed at constant middle forward speed maneuver for double lane change scenario through simulation. The result shows the predictive control based on Kautz function give a better tracking performance than conventional method, but has more complexes in term of algorithm implementation. It also demonstrates that Kautz function has a potential to eliminate the unmeasured disturbance or crosswind effect and it significantly improves the vehicle stability, and the trajectory along the desired path.