摘要
高速列车自动驾驶系统是一个快时变、非线性的复杂受控系统,针对运行阻力难以建模的问题,建立非参数化的阻力模型,提出基于径向基的模糊神经网络算法对其进行逼近。在此基础上,考虑高速列车自动驾驶系统具有高度重复性的特点,结合现有算法反馈控制的思想,引入前馈控制模型,设计非参数化的迭代学习控制律,充分学习系统的重复性信息,基于Lyapunov稳定性原理,经过严格的数学推导证明所提出算法的收敛性。同时,利用计算机对算法进行仿真试验和分析,验证所提出算法具有较快的收敛性和对期望曲线的较高跟踪精度。
High-speed train automatic train operation system is a fast time-varying,nonlinear and complicated control system.In order to solve the operational resistance modeling problem,the non-parametric resistance model was established and the fuzzy neural network algorithm based on radial basis function was proposed.Then according to the highly repetitive character of automatic train operation system,the feedforward control model was introduced based on the existing feedback control principle.The non-parametric iterative learning control law was designed to learn the repetitive information of the system.Based on Lyapunov Function,the convergence of the proposed algorithm was proved through strict mathematical referring.In the meantime,the computer simulation and analysis of the proposed algorithm show that the proposed algorithm features fast convergence and high tracking accuracy to the desired curve.
作者
何之煜
徐宁
HE Zhiyu;XU Ning(Signal & Communication Research Institute, China Academy of Railway Sciences Corporation Limited, Beijing 100081, China)
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2020年第12期90-96,共7页
Journal of the China Railway Society
基金
中国铁路总公司科技研究开发计划(2017X002,P2018G009)。
关键词
迭代学习控制
非参数化模型
复合能量函数
收敛性分析
自动驾驶
iterative learning control
non-parametric model
composite energy function
convergence analysis
automatic driving
