Research On BDS/INS Train Combination Positioning Information Fusion Algorithm Based On Improved CPF

With the rapid development of science and technology,positioning technology can be seen everywhere in our lives.A single positioning system cannot meet the user’s requirements for positioning accuracy,positioning reliability,availability and real-time performance,so the integrated navigation technology emerges as the times require.Global Navigation Satellite System(GNSS)and Inertial Navigation System(INS)can complement each other’s limitations and can significantly improve the overall performance of the navigation system.They are highly valued by scholars at domestic and abroad.It is also a research hotspot in the field of integrated navigation,and is of great significance to the optimization of the train control system.Based on this,this paper takes the train control system as the background,and combines the BDS and INS to improve the positioning accuracy of the system.The main contents are as follows:(1)In view of the urgent need to improve the positioning accuracy in the combined positioning system,it is necessary to select an appropriate filtering algorithm.In this paper,the CPF algorithm and several commonly used filtering algorithms are used to filter the nonlinear model.By comparison,it can be known that the filtering performance of the CPF algorithm is more superior.In order to further optimize the CPF algorithm and solve the problem that the system running rate is reduced due to excessive accumulation of useless particles during the resampling process,an ant colony optimization CPF algorithm(ACO-CPF)is adopted.The CPF and ACO-CPF algorithms are respectively applied to the simulation model of the train combined positioning system and simulated.The simulation results show that the ACO-CPF algorithm has higher filtering accuracy and has the potential to be applied in practical engineering.(2)Aiming at the problem that the BDS satellite signal will fail under extreme conditions during train operation and cannot meet the requirements of train positioning accuracy,a neural network is introduced into the BDS/INS train combined positioning model.When the BDS satellite signal fails,the LSTM1 network trained when the BDS satellite signal is valid and the INS information are used as the input of the ACO-CPF2 algorithm to realize the initial startup of the system when the BDS satellite signal fails;then the LSTM2 network integrates the current error,historical error,historical control amount,and use the information filtered by the ACO-CPF2 algorithm to learn.The simulation results show that in the case of BDS satellite signal failure,if no network assistance is added,the filtering effect will gradually diverge.Its positioning effect is ideal.