Study On Integrated Positioning Algorithm Of High Speed Train Based On The Neural Network

With the rapid development of China High Speed Railway,the train running speed is constantly improving,and the requirements for positioning accuracy are also constantly improving.However,the traditional train positioning method heavily depends on trackside equipment,and the positioning accuracy is low,which can’t meet the current positioning requirements of high-speed trains.Therefore,it is urgent for high-precision positioning methods to ensure the safe operation of high-speed trains.Considering the shortcomings of traditional positioning technology,in order to ensure the safety of high-speed railway train operation,based on the demand of next generation train control system in China,this thesis uses BDS/SINS integrated technology to realize the accurate positioning of the train.The integrated positioning technology can overcome the shortcomings of their own positioning,and have excellent positioning performance and can carry out safe and reliable train positioning.Therefore,based on the BDS / SINS integrated positioning technology,in order to improve the positioning accuracy and reliability of high-speed train,this thesis carried out the following research:Firstly,this thesis introduces the earth model,the earth’s basic parameters,different coordinate systems and the neural network,analyzes the structure and working principle of BDS and SINS system in detail,and fully studies the positioning error sources of BDS system and SINS system.Secondly,considering that KF algorithm needs to give the system model in advance and determine the statistical characteristics of noise,however,in actual process,the system model and the statistical characteristics of noise have time-varying characteristics,which are usually different from the predetermined model and the statistical characteristics of noise.To solve those problems,this thesis proposes ElmanKF algorithm based on Elman neural network,designs a tightly integrated train positioning system based on Elman-KF algorithm,and carries out positioning simulation experiments on KF algorithm,AKF algorithm and Elman-KF algorithm.The results show that Elman-KF has the highest positioning accuracy,which proves that Elman neural network can effectively make up for the shortcomings of KF algorithm.Finally,there may be BDS signal interruption in the process of high-speed train operation.For example,when the train is running in tunnels,forests,mountainous areas and other environments with weak BDS received signal,the train mainly realizes effective positioning through SINS system at this time.However,when SINS is used to locate alone,there is a problem that the positioning error is accumulating,which leads to the continuous decline of positioning accuracy.Therefore,this thesis proposes BDS/SINS tightly integrated technology based on RBF-Elman-KF algorithm.When the train can receive BDS signal normally,it realizes the accurate training of RBF neural network.When BDS signal is interrupted or abnormal signal reception,RBF is in the prediction state,in this stage,the trained RBF neural network is used to compensate the output of SINS.However,if the BDS signal is interrupted for a long time,the RBF prediction phase error will accumulate.This thesis solves the above problem by setting the key balise to correct the train position.The simulation results show that the tightly coupled positioning system based on RBF-Elman-KF algorithm and adding key balise can effectively suppress the cumulative positioning error of SINS system when BDS signal is interrupted.