交通运输系统工程与信息 ›› 2023, Vol. 23 ›› Issue (4): 134-146.DOI: 10.16097/j.cnki.1009-6744.2023.04.014

• 智能交通系统与信息技术 • 上一篇    下一篇

基于扩张状态观测器的虚拟编组触发模型预测控制

林俊亭*,倪铭君   

  1. 兰州交通大学,自动化与电气工程学院,兰州 730070
  • 收稿日期:2023-03-26 修回日期:2023-05-26 接受日期:2023-05-29 出版日期:2023-08-25 发布日期:2023-08-21
  • 作者简介:林俊亭(1978- ),男,河北海兴人,教授,博士
  • 基金资助:
    国家自然科学基金(52162050)

Trigger Model Predictive Control Based on Extended State Observers for Virtual Coupling

LIN Jun-ting*, NI Ming-jun   

  1. School of Automation and Electrical Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China
  • Received:2023-03-26 Revised:2023-05-26 Accepted:2023-05-29 Online:2023-08-25 Published:2023-08-21
  • Supported by:
    National Natural Science Foundation of China (52162050)

摘要: 为研究实际线路环境下的虚拟编组列车编队控制问题,综合考虑系统的实时性和抗扰性,本文提出一种基于扩张状态观测器的触发模型预测控制方法。首先,基于虚拟编组列车动力学方程搭建多车追踪模型,并设计终端函数,确保系统的稳定性;其次,针对模型预测控制中存在计算效率低和资源浪费的问题,引入事件触发机制,在求解优化问题时增加判断条件,以提高系统的实时性;再次,在前馈通路上加入扩张状态观测器,对环境中存在的扰动实时估计并补偿,减少干扰造成模型不准确的现象,提高模型的抗扰动能力;最后,基于MATLAB和Simulink仿真平台设计4列列车追踪的运行场景,并与传统模型预测控制方法进行对比,从控制精度和计算时间两方面验证所提方法的有效性。仿真结果表明:相较于传统模型预测控制方法,本文所提出的事件触发机制中3个阈值 τ为0.001,0.010,0.100 的计算效率分别提高了47%,64%,73%;在面对外部扰动时,本文所提算法在抗扰性上提高了33%。

关键词: 铁路运输, 列车编队控制, 模型预测控制, 虚拟编组列车组, 事件触发机制, 扩张状态观测器

Abstract: This paper proposes a trigger model prediction control method based on extended state observer to evaluate the real-time and resistance of the system for virtual coupling train control in the actual railway environment. First, a multi-train tracking model is developed based on the virtual coupled train dynamic equations, and the terminal function is designed to guarantee system stability. Then, to improve the system's real-time performance, this paper uses an event trigger mechanism to strengthen judgment conditions for the optimizations. This addresses the issue of low computational efficiency and resource waste in model prediction control. In addition, a front-end route extended state observer is introduced. This observer can estimate and correct the environmental disturbances in real time, decreasing the phenomenon of interference that leads to model inaccuracies and enhancing the model's resistance to distraction. To verify the effectiveness of the proposed method, four scenarios were built based on the MATLAB and Simulink simulation platforms and compared with conventional model prediction control methods. The results showed that compared to the traditional model prediction control methods, the three thresholds τ is 0.001,0.010,0.100 of the event trigger mechanisms increased the computational efficiency by 47% , 64% , and 73% respectively. The proposed algorithms increased the resistance by 33% in the conditions with external disturbances.

Key words: railway transportation, train platoon control, model predictive control, virtually coupled train set, event triggering mechanism, expansion state observer

中图分类号: