基于模型补偿的高速列车状态反馈预测控制

State feedback predictive control of high-speed train based on model compensation

高速列车运行系统本质上是高度非线性和不确定性的系统,为了弥补建模过程中被忽略或者简化的非线性和不确定性,提高高速列车运行过程的控制精度,提出一种基于模型补偿的高速列车状态反馈预测控制方法。在建模和控制上,分别采用子空间辨识法和状态反馈预测控制法,在此基础上建立BP神经网络在线补偿器,利用高速列车运行过程的状态变量和实际速度作为补偿器的输入,参考轨迹与实际速度之间的差值构成性能指标函数进行在线训练,输出补偿控制力作用于控制系统完成在线补偿,实现高速列车目标速度曲线高精度跟踪控制。仿真实验结果表明,该方法能够提高控制系统的控制精度。

The high-speed train operation system is essentially a highly nonlinear and uncertain system.In order to compensate for the non-linearity and uncertainty that are ignored or simplified during the modeling process,and to improve the control accuracy of the high-speed train operation process,method on model compensation of high-speed train state feedback predictive control is proposed.In modeling and control,the subspace identification method and the state feedback predictive control method are used respectively.On this basis,an BP neural network online compensator is established,and the state variables and actual speed of the high-speed train operation process are used as the input of the compensator.The difference between the trajectory and the actual speed constitutes a performance index function for on-line training,and the output compensation control force acts on the control system to complete the on-line compensation to achieve high-precision tracking control of the target speed curve of the high-speed train.Simulation experiment results show that this method can improve the control accuracy of the control system.