智能汽车轨迹跟踪MPC-RBF-SMC协同控制策略研究

Research on MPC-RBF-SMC Collaborative Control Strategy for Intelligent Vehicle Trajectory Tracking

针对自动驾驶车辆行驶过程中模型失配以及外部环境干扰导致车辆轨迹跟踪环节精确性不高的问题,提出了一种结合车辆运动学模型预测控制(MPC)、径向基(RBF)神经网络和滑模控制(SMC)的轨迹跟踪控制策略。通过建立车辆运动学MPC模型计算当前状态车辆期望横摆角速度,并将其与实际横摆角速度的偏差输入RBF-SMC控制器,利用RBF快速逼近非线性模型的特点,结合滑模控制输出前轮转角,实现车辆的横向轨迹跟踪控制。仿真结果表明,与传统的控制器相比,该方法轨迹跟踪精度显著提高,并在不同行驶工况下表现出较好的鲁棒性。

This paper proposed a trajectory tracking control strategy that combined Model Predictive Control(MPC),Radial Basis Function(RBF)neural network,and Sliding Mode Control(SMC)to address the low accuracy of vehicle trajectorytracking caused by model mismatch and external environmental interference during the driving process of autonomous vehicles.By establishing a vehicle kinematic model predictive control,the expected yaw rate of the vehicle in the current state wascalculated,and the deviation value from the actual yaw rate was input to the RBF-SMC controller.By utilizing RBF’s ability toquickly approach nonlinear models,combined with sliding mode control to output front wheel angles,the lateral trajectorytracking control of the vehicle was achieved.The simulation experimental results show that this method significantly improvestrajectory tracking accuracy compared with traditional controllers,and exhibits good robustness under different drivingconditions.