基于专家系统及在线调整的列车智能驾驶算法  

An Intelligent Train Operation(ITO)Algorithm Based on Expert System and Online Adjustment

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：《铁道学报》2014年第2期62-68,共7页Journal of the China Railway Society

基　　金：北京科技新星计划(2010B015);轨道交通控制与安全国家重点实验室自主课题(RCS2011ZT001);中央高校基本科研业务费专项资金(2012JBM016)

摘　　要：本文分析列车自动驾驶(ATO)系统中传统控制方法在灵活性和智能性方面的不足,结合驾驶经验和梯度下降法,提出基于专家系统及在线调整方法的列车智能驾驶(ITO)算法。总结操纵规则和驾驶经验,建立基于专家经验的控制策略以节约能耗、提高舒适度。针对列车运行控制的多目标性,在专家系统的基础上,引入基于梯度下降法的在线调整方法以减小运行时间和停车精度误差。在MATLAB软件环境下构建仿真模型,并运用实际线路进行仿真比较,与PID控制算法的对比结果表明,该算法优于PID控制算法,符合驾驶经验,可提高舒适度,降低能耗,满足停车精度要求,同时对不同的运行时间有较好的适应性,智能性高。The traditional control theory in the Automatic Train Operation （ATO） system lacks flexibility and intelligence .Combining the driving experience and gradient descent algorithm , the ITO algorithm based on the expert system and online adjustment methods was proposed for subway train control . Summarizing operat-ing rules and driving experience , the control strategy based on the experience of experts was built to reduce en-ergy consumption and better the degree of comfort . In view of the multi-objectives of train operation control , on the basis of the expert system , the online adjustment method based on the gradient descent algorithm was introduced to reduce running time and parking errors .By building an online simulation model in the MATLAB software environment and using actual line data , the algorithm was verified .Comparison with PID shows that the algorithm is better than PID . The algorithm is basically in line with the driving experience and it is able to upgrade comfort , reduce energy consumption , meet parking accuracy and attain good adaptability to different running time requirements and better intelligence .

关 键 词：城市轨道交通 专家系统 梯度下降法 MATLAB仿真 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] U283[自动化与计算机技术—计算机科学与技术]