Research On BDS/SINS/DR Train Combined Positioning Information Fusion Algorithm Based On SR-UIF

Safe and efficient train positioning has important practical significance for the development and construction of China’s railways.The current railway positioning method relies on a large number of ground positioning equipment to provide relevant real-time train information,but this positioning method has the disadvantages of high maintenance cost,equipment redundancy and low positioning accuracy,and is no longer suitable for the current railway operation environment.With the comprehensive completion of Chinese Bei Dou Satellite Navigation System,there are more and more calls for the use of satellite positioning in railway positioning.However,the complex train operating environment can easily lead to the loss of the satellite signal of the train,resulting in the divergence of positioning results.In this paper,a compact combined train positioning algorithm based on Square-Root Unscented Information Filter(SR-UIF)is introduced,which reduces the divergence of positioning results due to the accumulation of errors in the process of positioning result estimation.The combined positioning of BDS,SINS and DR establishes a multivariate combined positioning model based on the obtained positioning information,and then solves the position of the train.Therefore,a BDS/SINS/DR train combined positioning information fusion algorithm based on SR-UIF is proposed.The main research contents are as follows:(1)When the train can receive BDS information continuously and reliably,a train positioning information fusion algorithm based on Square-Root Unscented Information Filtering SR-UIF is introduced.Based on the characteristics of BDS/SINS closely coupled train positioning information fusion model,the algorithm first adopts the untraced information filtering algorithm,which effectively reduces the difficulty of filtering calculation while ensuring the filtering accuracy.Secondly,the algorithm introduces the root mean square filtering idea on the basis of the unscented information filtering algorithm,and performs square processing on the difference matrix,which effectively reduces the magnitude of the error in the variance matrix,thereby reducing the divergence phenomenon of the filtering results.The offline simulation results based on the measured train positioning data are analyzed,and the results show that the train positioning information fusion algorithm based on SR-UIF reduces the influence of noise on the train positioning system,reduces the position error and speed error of the train,and can still provide reliable positioning information for the train in a short time under the condition that the BDS signal is partially out of lock.(2)When the BDS signal is completely out of lock,dead reckoning DR is introduced to build a BDS/SINS/DR train combination positioning model,and the BDS/SINS combination system and DR/SINS combination system are used as sub-filtering systems to perform filtering processing in the sub-systems Finally,the fault detection process is performed on the local filtering results of the two subsystems,the filtering results with large errors are removed,and information fusion is carried out according to the effective probability of the subsystems,and finally the optimal estimation of the whole system is obtained.The experimental results show that when the BDS signal is completely out of lock,the BDS/SINS/DR train combined positioning information fusion algorithm based on SR-UIF can keep the positioning error within a relatively small range in a short time,and can effectively deal with the blind spot environment The following train positioning requirements.