自动泊车系统规划与控制方法研究

Research on planning and control method of automatic parking system

为了对泊车过程中的车辆进行跟踪控制,基于车辆动力学模型,研究设计了一种线性二次型控制器,使用B-spline曲线规划了一条曲率连续的泊车路径,针对控制器设计过程中权重系数选择困难,引入模糊逻辑算法对控制器进行优化处理,节省了人工成本,然而模糊逻辑算法的计算是实时的,且不同速度下车辆状态不同,因此将基于Frobenius范数的更新策略加入程序计算中以降低计算量,利用MATLAB/Carsim联合仿真验证了改进后控制系统的运算时间降为0.2 ms,实时性提升92%,适应度提升30%.

Based on the vehicle dynamics model,a linear quadratic controller is designed for the tracking control of the vehicle in the parking process,and a parking path with continuous curvature is planned using the B-spline curve.The selection of coefficients is difficult,and the fuzzy logic algorithm is introduced to optimize the controller,which saves labor costs.However,the calculation of the fuzzy logic algorithm is real-time,and the state of the vehicle is different at different speeds.Therefore,the update strategy based on the Frobenius norm is added to the program calculation in order to reduce the amount of calculation and the MATLAB/Carsim joint simulation is used to verify that the operation time of the improved control system is reduced to 0.2 ms,the real-time performance is increased by 92%,and the adaptability is increased by 30%.