Trajectory Tracking Control Based on Fuzzy-MPC for Autonomous Vehicles

In recent years, autonomous vehicles (AVs) have gained significant interests and it is expected that fully- autonomous vehicles will be appearing on roads soon. Trajectory tracking plays a crucial role in the driving tasks of autonomous vehicles. While classic model predictive control (MPC) is a common tracking method, it may not work well when encoun-tering sharp bends or curves. Aiming at solving this issue, this paper integrates fuzzy rules with classic MPC to design a new control method, namely, F-MPC. Fuzzy rules are set by considering curve radius, expected speed and steering angle of the reference trajectory. Then, the expected steering angle in the MPC model is adjusted according to the set fuzzy rules to achieve better tracking performance. In addition, the constraints are reasonably adaptive to the steering system of the actual vehicle. The effectiveness and advantages of the proposed trajectory tracking method have been verified through simulations and experiments on the CarSim-MATLAB/Simulink environment and QCar experimental platform in tracking S-shape trajectory.