城轨ATO系统子空间预测控制方法

颜争，兰清群

安徽交通职业技术学院，安徽 合肥  230051

摘要：为提高城市轨道交通列车自动驾驶（automatic train operation，ATO）系统跟踪给定运行曲线的精度，基于子空间辨识方法，利用列车运行的历史数据，建立与实际运行状态相吻合的非线性子空间预测控制模型，设计子空间预测控制器，实现模型辨识数据和参数在线更新。运用MATLAB软件对比分析传统动力学模型与子空间预测控制模型的跟踪能力。结果表明：子空间预测控制模型在速度、位移、加速度的跟踪精度上有明显优势，牵引/制动特性更加缓和。子空间预测控制模型可以保证列车运行安全、准时，并提高乘客乘坐舒适性。

关键词：城轨列车；ATO系统；子空间；预测控制；MATLAB

Subspace predictive control methods ofurban rail transit for ATO system

YAN Zheng，LAN Qingqun

Anhui Communications Vocational & Technical College，Hefei 230051，China

Abstract:In order to enhance the accuracy of automatic train operation (ATO) system aligned with the given operational curve, based on the subspace identification method and using the historical data of train operation, the nonlinear subspace predictive control model which is consistent with the actual running status is established, and the subspace predictive controller is designed. The model identification data and parameters are updated online. The tracking ability of traditional dynamics model and subspace predictive control model are analyzed by MATLAB. The results show that the subspace predictive control model has obvious advantages in terms of tracking accuracy of velocity, displacement, acceleration; and the traction/braking characteristics performs smoothly. The subspace predictive control model can ensure the safety and punctuality of train operation, then improve the passenger comfort.

Keywords:urban railway train; ATO system; subspace; predictive control; MATLAB