Research On Satellite Positioning Availability Evaluation And Prediction Method In Railway Scenario

Accurate train localization plays an important role for the safe train operation.With the development of Global Navigation Satellite System(GNSS),the Next Generation Train Control System(NGTC)and satellite-based train localization have become hot topics in many countries.GNSS signals are easily blocked by environmental occlusion,due to the complexity and the variety of the railway environment scenario,the GNSS availability changes during the journey of the train,and even lead to no location results in some cases.Therefore,evaluating and predicting the satellite positioning availability in different railway scenarios is significant to ensure the safe train operation.This thesis proposes a method for evaluating and predicting the satellite positioning availability in railway scenarios.Firstly,the scenarios along railway are classified by hierarchical clustering algorithm,environmental occlusion in each scenario are rebuilt,and GNSS availability in each scenario is evaluated.Then,the satellite geometry is predicted using satellite almanac and environmental occlusion.GNSS availability in each scenario is predicted based on Xgboost model and satellite geometry result.The contributions of this thesis are as follows:(1)Investigating the state of the art on GNSS and satellite-based train localization availability evaluation.Based on previous research results,this thesis proposes GNSS availability evaluation metrics in railway scenario based on SIS availability,HDOP availability,precision along track availability and precision vertical track availability.And divides the GNSS availability into three status:available,degraded and unavailable.(2)This thesis proposes a classification algorithm based on hierarchical clustering for railway scenarios.To extract the environmental characteristics,the environmental-related parameters such as visible satellites and satellite geometry from GNSS receiver logged raw data are preprocessed,then these characteristics are inputs of the hierarchical clustering ’algorithm.The scenarios along railway are classified according to the relationship between hierarchical clustering distance and steps.According to the classification results,the environmental occlusion on the satellite signal in each scenario is rebuilt,the GNSS positioning error and availability are also evaluated.(3)With the satellite almanac and environmental occlusion,the visible satellite number,satellite geometry,SIS/HODP availability in each environmental scenario are predicted.Taking the satellites geometric as inputs,using optimized Xgboost algorithm to predict precision along track availability and precision vertical track availability.Finally,GNSS precision availability is predicted.Based on the test runs along Qinghai-Tibet Railway,the GNSS availability evaluation and prediction methods in railway scenarios are verified in this thesis.The data analysis platform of Qinghai-Tibet railway is built to demonstrate visible satellites,HDOP value and the predicted GNSS availability along the railway.The test results show that the algorithm based on hierarchical clustering proposed in this thesis classifies the railway scenario as open,cut slope,tunnel,mountain and half-sky,and rebuilt the environmental occlusion.Comparing with 3D modeling and fish-eye method,it is proved that the method proposed can rebuilt environmental occlusion accurately.According to the environmental occlusion,using satellite almanac and Xgboost model to predicgt GNSS availability along railway,and its accuracy is above 95%.Figure 66,table 33,reference 66.