计算机工程与应用 ›› 2021, Vol. 57 ›› Issue (10): 252-258.DOI: 10.3778/j.issn.1002-8331.2002-0231

• 工程与应用 • 上一篇    下一篇

高速列车速度跟踪神经网络PID控制器的设计

梁新荣,肖龙,王雪奇,杨世武,董海荣   

  1. 1.五邑大学 交通工程系,广东 江门 529020
    2.北京交通大学 电子信息工程学院,北京 100044
  • 出版日期:2021-05-15 发布日期:2021-05-10

Design of Neural Network PID Controller for Speed Tracking of High-Speed Train

LIANG Xinrong, XIAO Long, WANG Xueqi, YANG Shiwu, DONG Hairong   

  1. 1.Department of Traffic Engineering, Wuyi University, Jiangmen, Guangdong 529020, China
    2.School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing 100044, China
  • Online:2021-05-15 Published:2021-05-10

摘要:

高速列车速度跟踪控制系统是一个复杂的非线性系统,难以取得高精度的跟踪性能。为了减少速度跟踪误差,设计了高速列车神经网络PID控制器。首先建立了描述列车运行过程的单位移多质点模型,该模型考虑了列车的基本阻力和附加阻力以及车厢之间的相互作用力。然后阐述了BP神经网络PID控制,并设计了列车速度跟踪控制器,根据速度误差用神经网络PID控制决定牵引力和制动力。最后与模糊控制和常规PID控制进行了仿真对比,结果表明,神经网络PID控制具有很小的速度跟踪误差和优越的速度跟踪性能,可以满足列车正点运行的需求。

关键词: 高速列车, 多质点模型, 速度跟踪, 神经网络, 反馈控制

Abstract:

The speed tracking control system for high-speed trains is a complex nonlinear system, and it is hard to achieve high-precision tracking performance. A neural network PID(Proportion-Integration-Differentiation) controller for the high-speed train is designed to reduce the speed tracking error. Firstly, a unit-displacement multi-particle model is built to describe the train running process. In this model, the basic resistance and extra resistance to the train, as well as the interaction between carriages, are considered. Then, PID control regulated by BP(Back Propagation) neural network is formulated, and a speed tracking controller for the train is designed. The traction force and braking force are determined by the neural network PID control based on the speed error. Finally, simulation comparison with a fuzzy control and a conventional PID control is accomplished. Simulation results indicate that the neural network PID control has a very small speed tracking error and has superior speed tracking performance. This method can meet the demand for the accurate running of high-speed trains.

Key words: high-speed train, multi-particle model, speed tracking, neural network, feedback control