计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (22): 224-229.DOI: 10.3778/j.issn.1002-8331.2005-0254

• 工程与应用 • 上一篇    下一篇

ATO系统速度控制的BP-FIPID算法

楚彭子,虞翊,林辉,袁建军,姜西   

  1. 1.同济大学 道路与交通工程教育部重点实验室,上海 201804
    2.同济大学 磁浮交通工程技术研究中心,上海 201804
    3.同济大学 上海市磁浮与轨道交通协同创新中心,上海 201804
    4.中铁第四勘察设计院集团有限公司,武汉 430063
  • 出版日期:2020-11-15 发布日期:2020-11-13

BP-FIPID Algorithm for Speed Control of ATO System

CHU Pengzi, YU Yi, LIN Hui, YUAN Jianjun, JIANG Xi   

  1. 1.The Key Laboratory of Road and Traffic Engineering, Ministry of Education, Tongji University, Shanghai 201804, China
    2.Maglev Transportation Engineering R&D Centre, Tongji University, Shanghai 201804, China
    3.Shanghai Maglev and Rail Transit Collaborative Innovation Centre, Tongji University, Shanghai 201804, China
    4.China Railway Siyuan Survey and Design Group Co., Ltd., Wuhan 430063, China
  • Online:2020-11-15 Published:2020-11-13

摘要:

针对列车自动运行(Automatic Train Operation,ATO)系统控制算法的稳定性与智能性需求,以及比例积分微分(Proportion Integration Differentiation,PID)控制算法的拓展优化,结合BP(Back Propagation)神经网络算法和模糊免疫PID(Fuzzy Immune PID,FIPID)控制算法,提出一种基于BP神经网络的免疫控制参数自适应调整的模糊免疫PID控制算法(BP-FIPID)。以列车运行控制模型为控制对象,分别采用阶跃信号和列车运行目标速度曲线对传统FIPID以及BP-FIPID进行仿真检验。测试结果显示,与FIPID算法相比,BP-FIPID算法具有更好的阶跃响应和抗干扰性能,针对复杂工况的速度-时间曲线同样体现出理想的追溯性。免疫控制参数的自适应调整有助于改进FIPID的性能,两种算法均可作为实践参考。

关键词: 列车运行控制, ATO系统, BP-FIPID, 模糊免疫PID(FIPID), BP神经网络

Abstract:

Aiming at the stability and intelligence requirements of the control algorithm in Automatic Train Operation(ATO) systems, and the expansion and optimization of PID control algorithm, a BP (Back Propagatio) neural network-based Fuzzy Immune PID(BP-FIPID) control algorithm for the adaptive adjustment of immune parameters is proposed based on the BP neural network algorithm and Fuzzy Immune PID(FIPID) control algorithm. Taking a train operation control model as the control object, FIPID and BP-FIPID are tested using a step signal and a target speed curve of train operation. The results show that compared with FIPID, BP-FIPID algorithm has a good performance of step response and anti-interference, and the traceability for the target speed curve with complex conditions is also ideal. The adaptive adjustment of immune control parameters can help to improve the performance of FIPID, and the two algorithms can be used as practical references.

Key words: train operation control, ATO system, BP-FIPID, Fuzzy Immune PID(FIPID), BP neural network