摘要
为在列车运动过程非线性变化、噪声为非高斯的条件下,实现测速传感器的量测误差校准及融合后的速度高精度估计,提出列车测速系统配置双雷达与轮轴传感器,并根据联邦粒子滤波进行融合的测速方法。以CRH3型列车在两车站间从启动到停止的运行全过程为例,将不同速度估计方法在列车产生空转、打滑、振动以及测速传感器存在动态噪声情况下的测速误差进行分析并验证。仿真结果表明:双雷达与轮轴组合进行校准后粒子滤波相对未校准时在空转和打滑阶段均方根误差可分别降低31.52%、47.35%;对比双雷达分离振动速度校准法和双雷达角度偏差估计校准法,双雷达分离振动速度校准法在列车振动速度占比0~1和雷达安装角误差-1°~1°范围内,相对误差最大可降低39.66%。采用联邦粒子滤波融合后的测速结果较融合前双雷达和轮轴传感器滤波测速的结果MAE分别降低34.71%,14.03%,RMSE分别降低26.51%、10.98%。采用联邦粒子滤波融合的方式较联邦扩展卡尔曼滤波融合测速均方根误差降低26.97%,最大绝对误差可降低16.10%。
For calibrating the measurement error of the speed sensor and estimating the speed accurately while system is nonlinear and noise is non-Gaussian,dual radar and wheel axle sensor are configured and federal particle filter is used to acquire fusion.The whole operation process of CRH3 train from start to stop between two stations is taken as an example,the speed measurement errors of different speed estimation methods under the conditions of train idling,slipping,vibration and the speed sensor with dynamic noise are analyzed and verified.The simulation results show that:The root-mean-square error can be reduced by 31.52%and 47.35%respectively in the idling and sliding stages after the calibration of dual-radar and wheel-axle combination compared with no calibration;Compared with the two types of dual-radar calibration methods,the maximum relative error of the dual-radar separation vibration speed calibration method can be 39.66%lower than that of the dual-radar angle deviation estimation calibration method when the train vibration speed ratio is 0 to 1 and the radar installation angle error is-1°to 1°.Compared with the filtering results of dual radar and axle sensors before fusion,the speed measurement results after using joint particle filtering fusion are 34.71%and 14.03%lower in MAE,and 26.51%and 10.98%lower in RMSE respectively.Compared with federated extended Kalman filter fusion,the root mean square error of velocity measurement using federated particle filter fusion is 26.97%lower and the maximum absolute error can be reduced by 16.10%.
作者
李泽鑫
张亚东
魏维伟
何静
王小敏
LI Zexin;ZHANG Yadong;WEI Weiwei;HE Jing;WANG Xiaomin(School of Information Science and Technology,Southwest Jiaotong University,Chengdu 611756,China;Sichuan Province Train Operation Control Technology Engineering Research Center,Chengdu 611756,China;Traffic Sensing Radar Research and Development Center,China Aerospace Science and Technology Corporation,Shanghai 201109,China)
出处
《铁道标准设计》
北大核心
2024年第6期221-228,共8页
Railway Standard Design
基金
中国国家铁路集团公司科技研究开发计划项目(N2021G045,N2021T008,P2021G053)
上海航天科技创新基金项目(SAST2020-126)。
