The adaptive MPC controller of autonomous driving formula for trajectory tracking

Autonomous driving technology has gradually been applied in real life, and it has also been improved in the China Collegiate Electric Vehicle Racing. However, as one of the critical components of autonomous driving technology, trajectory tracking control still faces challenges. This paper designs a trajectory tracking controller by building a vehicle's three-degree-of-freedom dynamics model and adopting adaptive model predictive control. It adjusts the prediction horizon of the vehicle's trajectory tracking controller in real-time based on the curvature of the desired trajectory to improve the effect of trajectory tracking. Experimental results show that the tracking effect of the adaptive prediction time-domain trajectory tracking controller in the S-curve is improved by 4.97% and 11.36% compared to the fixed prediction time-domain trajectory tracking controller, respectively. For the formula track conditions, the tracking effect of the controller proposed in this paper is improved by 0.71% compared to the fixed prediction time-domain trajectory tracking controller, with an average tracking error of 5.65%.