铁道科学与工程学报2025，Vol.22Issue(4)：1492-1505,14.DOI:10.19713/j.cnki.43-1423/u.T20241028
基于惯性四元数建模的列车姿态精度保持方法研究
Research on train attitude accuracy maintenance method based on inertial quaternion modelling
摘要
Abstract
To address the issue of error accumulation in inertial devices and the resulting degradation of train state perception accuracy when the standard GNSS/INS positioning model fails due to satellite failure,and to ensure the spatial accuracy and robustness of the train in the event of satellite failure,a train attitude accuracy maintenance method based on inertial quaternion modelling was proposed.The attitude solution mechanism of acceleration and angular velocity of inertial devices was explored,and a feedback compensation solution model for inertial attitude quaternion assisted by the motion carrier's acceleration was constructed.Further,to solve the problem of limited controllability of the single drive mode on the heading estimation,a train state perception model was established based on the enhancement of the attitude accuracy maintenance model.The attitude information fused and solved in the inner layer was mapped to the outer layer state space with the inertial recursive value to complete the system's filter correction and fed back to the inner layer for error updating,thereby maximizing the protection of inertial navigation accuracy during satellite failure.The algorithm performance was verified and evaluated using simulation data from the laboratory turntable and measured data from the Qinghai-Tibet Line train.The results show that the proposed attitude accuracy maintenance method was significantly superior to the inertial recursive method.In the 30-minute dynamic posture measurement on the Qinghai-Tibet Line,the roll,pitch,and heading angle accuracy were 0.801°,0.513°,and 4.494°,respectively.In the absence of satellite scenarios,the proposed train state perception method achieved a roll,pitch,and heading angle measurement accuracy of 0.526°,0.493°,and 1.904° during the 15-minute real-train test,respectively,which were significantly better than those of the INS/ODO loose combination method.The horizontal positioning accuracy was improved by 86.66%,and the horizontal speed measurement accuracy was improved by 78.70%.These results demonstrate that this method contributes to improving the inertia-based train state perception accuracy and robustness in satellite failure scenarios.关键词
惯性四元数建模/姿态精度保持/惯性递推/滤波校正/列车状态感知Key words
inertial quaternion modelling/attitude accuracy maintenance/inertial recursive/filter correction/train state perception分类
交通运输引用本文复制引用
姜维,郝鹏琪,王剑,蔡伯根,邓陈喜,廖浩..基于惯性四元数建模的列车姿态精度保持方法研究[J].铁道科学与工程学报,2025,22(4):1492-1505,14.基金项目
国家自然科学基金资助项目(T2422002) （T2422002）
国家重点研发计划(2022YFB4300501) （2022YFB4300501）
国家自然科学基金重大科研仪器研制项目(62027809) （62027809）
