Aiming at shortcomings of low positioning precision and difficulty of railway practical application

in railway scenario of existing cellular network localization technology

propose a new method of train positioning. The core of method is to process positioning parameters of speed sensor and cellular network

as possible position area of several trains and calculate probability of position of train in area

then use the maximum probability in all the positions as optimization objective

error optimization model is established for train positioning. The genetic particle swarm optimization（ PSO） algorithm is introduced to solve the model and simulate by software Matlab. The simulation results show that in complex environment of railway

the localization method has good absorbing effect on errors of different sources and it is an effective positioning method.