Research On Train Positioning Method Based On Risk Assessment

In the rail transit system,the determination of the train position is an important part of the train operation control system,and it is also a prerequisite for ensuring the safety of trains and improving the efficiency of operation.At present,the information output by a single positioning method cannot satisfy the needs of the new train control system.In order to better obtain train position information,it is necessary to combine multiple positioning methods with complementary performance.GPS(Global Positioning System)/SINS(Strapdown Intertial Navigation System)train positioning technology can meet the needs of new train control systems,and it is an inevitable trend in the development of train positioning in the future.However,the system has not been used for a long time in the field of railways in my country,and many risks have not been fully exposed.Therefore,potential risks should be analyzed to improve the safety of the system.In addition,it is also necessary to propose relevant improvement strategies based on the estimated greater influential factors.Aiming at the problem that the GPS signal is invalid caused by the obstruction of foreign objects,which affects the positioning results,a new positioning method is proposed to ensure the accuracy of the output information even if the GPS signal is invalid.Firstly,the positioning system of SINS,GPS and the combination of the two are briefly introduced.Based on the train combined positioning system,its internal structure of the system board is analyzed to lay the foundation for the establishment of the evaluation index system in the subsequent system evaluation.And the error source of GPS/SINS train positioning system is analyzed to provide a basic theory for the establishment of mathematical models in the next research on the positioning method.Secondly,D-AHP(D Number Optimization Analytic Hierarchy Process)and TOPSIS(Technique for Order Preference by Similarity to an Ideal Solution)are used to evaluate the GPS/SINS train positioning system.Taking the two out of two train positioning system as an example,the impact of system component failures and environmental factors on the system are analyzed to determine the evaluation index system.D-AHP and TOPSIS methods are used to solve the weights of various factors,expert weights and system risk quantification values,so as to determine the higher influential factors and the risk level of the system.Finally,according to the obtained risk level,it is determined that the system is necessary to be improved,and the larger influential factor is foreign object occlusion,which is determined based on the calculated weight.Aiming at the influential factor of foreign object occlusion,a solution of LSTM(Long Short Term Memory Networks)aided UPF(Unscented Particle Filter)is proposed.Based on the analysis of error sources,error mathematical models are established,and the positioning model when the GPS signal is valid and the positioning model when the GPS signal is invalid are established.The positioning results when the GPS signal is valid,without the algorithm aided UPF when the GPS signal is invalid and with the LSTM aided UPF when the GPS signal is invalid are compared,so as to show the necessity of LSTM aided UPF after GPS signal outages and the feasibility of LSTM aided UPF.