铁道标准设计2026，Vol.70Issue(3)：190-197,8.DOI:10.13238/j.issn.1004-2954.202404190007
侧风环境下城轨列车自动驾驶控制策略研究
Control Strategy for Automatic Train Operation of Urban Rail Trains Under Crosswind Conditions
摘要
Abstract
This paper primarily investigated automatic train operation in urban rail transit systems,with a particular focus on energy-saving and safety control strategies under crosswind conditions.It detailed the construction of the train dynamics model,multi-objective optimization modeling,and impacts of crosswind conditions on operation.An automatic train operation strategy based on Deep Reinforcement Learning(DRL)was proposed.The sparse reward problem was addressed through Tr(Time redundancy)reference planning,and a detailed reward function was designed to guide the agent's learning process.In addition,the DQN-Tr algorithm combined Tr reference planning with Deep Q-Network(DQN),improving the convergence speed and stability of the algorithm.Experimental simulations were conducted using a subway line in Chongqing as a case study,and the approach was validated.The results showed that the DQN-Tr algorithm performed well in stopping accuracy and punctuality,while achieving significant energy savings.Compared with traditional PID control,DQN-Tr increased accuracy by 0.05％and reduced energy consumption by 11.45％,effectively lowering energy use while ensuring operational safety,thereby providing an efficient and energy-saving control method for urban rail transit.关键词
城市轨道交通/列车自动驾驶系统/节能优化/深度强化学习/Tr参考规划/动力学模型/多目标优化/侧风环境Key words
urban rail transit/automatic train operation system/energy-saving optimization/deep reinforcement learning/Tr reference planning/dynamics model/multi-objective optimization/crosswind conditions分类
交通工程引用本文复制引用
杨浩博,陈晓强,郭佑民,胥如迅,李德仓..侧风环境下城轨列车自动驾驶控制策略研究[J].铁道标准设计,2026,70(3):190-197,8.基金项目
国家自然科学基金项目(72061021) （72061021）
甘肃省科技计划项目(22JR11RA146) （22JR11RA146）
兰州交通大学青年基金资助项目(2021018) （2021018）
