摘要
针对使用微分对策进行列车避碰分析时所存在的复杂计算问题,提出基于自适应神经网络的微分对策求解方法,分析列车运行过程中存在的冲突问题,进而避免列车碰撞事件的发生.将微分对策求解时遇到的双边极值问题转化为神经网络的学习问题,利用微分对策理论建立两追踪列车之间的追逃对策模型,并在Simulink中搭建仿真环境,对不同情况下列车追踪运行情况进行仿真验证分析.仿真结果表明:自适应神经网络能有效解决微分对策在分析列车避碰时的双边极值问题,为不同情况下快速合理地分析列车碰撞防护提供了一定的理论支撑.
In view of the complex calculation problems in the analysis of train collision avoidance by using differential game, a solution of differential game based on adaptive neural network is proposed to analyze the conflicts in the process of train operation, so as to avoid the occurrence of train collision.By transforming the bilateral extreme value problem encountered in differential game solution into a neural network learning problem, the differential game theory is used to establish the pursuit game model between two tracking trains, and the simulation environment is built in Simulink to verify and analyze the train tracking operation under different conditions The simulation results show that the adaptive neural network can effectively solve the bilateral extreme value problem of differential game in the analysis of train collision avoidance, which provides a theoretical basis for the rapid and reasonable analysis of train collision avoidance protection under different conditions.
作者
林俊亭
闵晓琴
王海斌
梁化典
LIN Jun-ting;MIN Xiao-qin;WANG Hai-bin;LIANG Hua-dian(School of Automation and Electrical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China;AVIC Zhonghang Electronic Measuring Instrument Co.,Ltd.,Xian 710119,China)
出处
《兰州交通大学学报》
CAS
2022年第5期28-33,共6页
Journal of Lanzhou Jiaotong University
基金
国家自然科学基金(52162050)
中国国家铁路集团有限公司科技研究开发计划项目(N2021G045)。
关键词
微分对策理论
列车碰撞防护
自适应神经网络控制器
双边极值
differential game theory
train collision protection
adaptive neural network controller
bilateral extreme value
