Research On Railway Environmental Scenario Modeling For GNSS-based Train Localization Performance Optimization

With the development of train operation control system,whether the train positioning performance based on global navigation satellite system(GNSS)can meet the needs of train control in the whole scenario is a hot research direction.The environment along the railway is complex and changeable.The sheltered environment scenarios formed by mountains,buildings and forests on both sides of the track will block and reflect the transmission of satellite signals,resulting in varying degrees of degradation of satellite positioning performance.In this paper,a satellite positioning accuracy optimization method based on railway environment scenario modeling is proposed.Firstly,the railway line and the environment scenario along the line are parameterized,and the three-dimensional digital track map(3D digital track map,3D DTM)of the railway environment scenario is constructed;Aiming at the general occlusion environment scenario,the weighted least square method and 3D DTM are used to optimize the satellite positioning accuracy;For severely occluded environment scenarios,the shadow matching algorithm is used to realize effective satellite positioning based on 3D DTM,and the satellite positioning accuracy is optimized combined with Kalman filter algorithm.The contributions of this thesis are as follows:(1)From the perspective of satellite positioning,the 3D DTM architecture of railway environment scenario is designed,including geographic information data,fixed application data and sky occlusion boundary data.Based on the data requirements of sky occlusion boundary in 3D DTM,an environmental scenario identification algorithm along the railway based on dynamic time warping(DTW)algorithm is proposed to determine the sky occlusion degree of environmental scenarios along the railway.(2)For general occluded environment scenarios(such as cutting,etc.),the recognition methods of line of sight(LOS)signal and non line of sight(NLOS)signal are studied.Through the application strategy of NLOS signal,the satellite positioning accuracy is optimized.When the NLOS signal is involved in the solution,the conditional probability analysis of Los / NLOS signal is carried out based on Bayesian method,the conditional probability table(CPT)of signal to noise ratio(SNR)is constructed,and the weighted least square method based on CPT is proposed;When the NLOS signal does not participate in the solution,the NLOS signal is eliminated through the sky occlusion boundary in 3D DTM,and the least square method is used for positioning solution.(3)Aiming at the severely occluded environment scenarios(such as tunnel entrance,deep urban canyon,etc.),the shadow matching algorithm based on 3D DTM is studied to optimize the satellite positioning accuracy,and the Kalman filter algorithm is used to fuse the speed information and the shadow to match the train position,so as to realize the satellite positioning accuracy optimization based on the improved shadow matching algorithm.This paper uses the measured satellite data of Beijing Sanjiadian station to verify the proposed algorithm.The experimental results show that the error rates of the proposed scenario identification algorithm for five types of railway environment scenarios(S1-S5)are 0.92%,1.84%,0.92%,0.00% and 0.00% respectively;In general occluded environment,compared with the traditional least square algorithm,the weighted least square method based on CPT and 3D DTM assisted positioning optimization method reduce the satellite positioning error from 5.09 m to 4.23 m and1.77 m respectively;In the severely occluded environment,compared with the output results of the receiver,the shadow matching algorithm and the shadow matching algorithm based on Kalman filter reduce the satellite positioning error from 14.81 m to3.34 m and 1.05 m respectively.Figures 66,tables 29,references 70.