基于纵横向多速度融合的高速列车测速精度研究  

Research on speed measurement accuracy of high-speed train based on longitudinal and transverse multi-speed fusion

作　　者：侯涛[1] 赵廷阳 Hou Tao;Zhao Tingyang(School of Automation and Electrical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学自动化与电气工程学院,兰州730070

出　　处：《北京交通大学学报》2023年第5期48-55,共8页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：甘肃省重点研发计划(23YFGA0049);甘肃省自然科学基金(21JR7RA321,22JR5RA358)。

摘　　要：针对现有高速列车速度控制多数存在的测速误差大、运行效率偏低等问题,提出一种纵横向多速度融合测速方法.首先,通过叠加式采样读取4路速度传感器的测速数据,采用联邦卡尔曼滤波算法对4路测速值分别进行纵向融合滤波,并采用衰减记忆法优化该算法存在的滤波发散问题,得到4路纵向融合速度值.其次,对4路纵向融合速度值进行置信距离可靠性判别,确定有效融合速度值数量,排除传感器故障影响.再次,采用优化的贝叶斯数据融合算法对有效纵向融合速度值进行横向多速度融合.最后,对所提算法进行仿真,完成对仿真结果的分析和比较.研究结果表明:基于衰减记忆法的联邦卡尔曼滤波算法的纵向融合速度平均误差为0.6696 km/h,纵横向多速度融合方法的融合速度平均误差为0.3928 km/h,平均测速精度有较大的提升.In view of the common issues of significant speed measurement errors and low operational efficiency in high-speed train systems,this study introduces a speed measurement method based on longitudinal and transverse multi-speed fusion.Firstly,the process begins by collecting speed data from four speed sensors through stacked sampling.The Federal Kalman filtering algorithm is applied to filter each of the four speed values longitudinally.The decay memory method is incorporated to ad-dress filtering dispersion issues,obtaining the 4 longitudinal fused speed values.Secondly,a confi-dence distance reliability of the four longitudinal fusion speed values is used to determine the number of valid fusion speed values,eliminating the impact of sensor failure.Thirdly,an improved Bayesian data fusion algorithm is employed to transversely fuse the valid longitudinal fused speed values.Finally,the algorithms are simulated,and the analysis and comparison of the simulation results are completed.The results show that the average error in the longitudinal fusion speed,when using the Federal Kalman filtering algorithm based on the decay memory method,is 0.6696 km/h.On the other hand,the average error in the fusion speed,utilizing the longitudinal and transverse multi-speed fusion method,is 0.3928 km/h,marking a substantial enhancement in average speed measurement accuracy.

关 键 词：测速精度 纵横向多速度融合 联邦卡尔曼滤波 衰减记忆法 贝叶斯数据融合 

分 类 号：U271.91[机械工程—车辆工程]