基于数据驱动的列车智能驾驶算法研究  

Study on intelligent train operation(ITO) algorithm based on data-driven

作　　者：冷勇林 陈德旺[1] 阴佳腾 

机构地区：[1]北京交通大学轨道交通控制与安全国家重点实验室,北京100044

出　　处：《铁路计算机应用》2013年第10期1-4,共4页Railway Computer Application

基　　金：教育部基本科研业务费(2012JBM016);北京市科技新星项目(2010B015);国家863计划(2011AA10502)

摘　　要：结合数据驱动控制理论和司机驾驶经验,提出数据驱动推理控制算法。考虑到列车运行控制的复杂性,在推理控制算法的基础上引入了牛顿迭代学习法以减小运行误差。在MATLAB环境构建了在线仿真模型,并运用实际线路进行了仿真比较,与PID控制算法的结果对比表明,该算法优于PID控制算法,基本符合驾驶经验,提高了舒适度、降低了能耗,同时满足停车精度和运行时间要求。Combined with the driving experience, an inference control algorithm based on data-driven control theory was proposed. Taking account of the complexity of train operation control, an online adjustment method based on the Newton Method was added to reduce operational errors. By building the online simulation model in the MATLAB software environment and the use of actual circuit data, the algorithm was verified. The results of comparing with PID showed that the algorithm was better than PID, basically in line with the driving experience and improves comfort, reduced energy consumption, met the parking accuracy and running time.

关 键 词：城市轨道交通 ATO 数据驱动 牛顿迭代法 PID 

分 类 号：U284[交通运输工程—交通信息工程及控制] TP39[交通运输工程—道路与铁道工程]