自适应时域模型预测控制的轨迹跟踪控制OA北大核心CSTPCD

To improve the trajectory tracking accuracy and stability of the self-driving vehicle, and the adaptive ability of the controller in different working conditions, this paper proposes a fuzzy adaptive prediction time domain parameter model predictive control trajectory tracking control algorithm.Firstly, the vehicle kinematics model and model predictive controller are built.Secondly, the real-time vehicle speed and heading angle deviation are taken as fuzzy inputs.The predictive time-domain parameters of MPC are optimized online through fuzzy control.The trajectory tracking control simulations with different vehicle speeds are conducted on asphalt pavement and rainy and snowy pavement through the joint simulation of Carsim/Simulink respectively.Our simulation results show compared with the fixed prediction time domain parameter MPC controller, the mean value of the longitudinal absolute deviation of adaptive time domain MPC controller reduces by 63.16％ and 55.28％ at low and high speeds respectively, effectively improving the trajectory tracking accuracy of vehicles on asphalt road.Meanwhile, the adaptive time domain controller' s maximum sideslip angle is within 1° at low or medium-high speeds under rainy and snowy road surfaces, ensuring the robustness of vehicle control.