Research On Positioning System Based On BDS-INS Integrated Navigation

With the advent of the intelligent era,all kinds of emerging industries,such as intelligent vehicles and robots,have an increasing demand for navigation and positioning technology.At present,the Beidou navigation satellite system(BDS)independently built in China has covered the world and can provide users with all-weather navigation and positioning services.However,it is vulnerable to the interference of the external dynamic environment in work,resulting in the deterioration or even loss of satellite signal quality.Inertial navigation system(INS)can provide continuous navigation and positioning information due to its autonomy,but its accuracy will diverge with time accumulation.Combining the two subsystems and learning from each other can greatly improve positioning accuracy.Therefore,the combined system based on the two has attracted extensive attention and become a research hotspot in this field.To solve the above problems,we study the Beidou and inertial navigation integrated navigation and positioning system.The main work is as follows:Firstly,the coordinate system and coordinate transformation used in the integrated system are introduced,the positioning principles and error sources of the Beidou satellite navigation system and inertial navigation system are explained,the error model of the system is constructed and analyzed by the Allan variance method.The results show that Allan variance method can effectively identify the error components and parameter indexes of inertial measurement instruments.Secondly,three combination modes of the two subsystems are introduced.Aiming at the problems of the standard Kalman filter in the traditional combination system,the strong tracking Kalman filter(STKF)is introduced.On this basis,the loose combination mode is selected,the 15-dimensional system state-space equation is established,and the vehicle driving process is simulated by simulation experiment.The performance of the two filtering algorithms is compared,which proves the superiority of the strong tracking Kalman filter algorithm in performance.Compared with the standard Kalman filter algorithm,the RMSE of the strong tracking Kalman filter algorithm in the east direction is reduced by about 10.38% and 52.48% respectively,and the RMSE of the north direction is reduced by 14.21% and 51.47% respectively.Then,aiming at the short-term loss of Beidou signal lock in a complex environment,an improved extreme learning machine neural network(IELM)assisted integrated navigation and positioning system is designed.When the Beidou signal is valid,the inertial navigation data is used as the network input,and the error information output by STKF filter is used as the network expected output to train the network;When the Beidou signal loses lock temporarily,the trained network prediction error information is used to correct the output of inertial navigation system.Finally,the software and hardware platform of the combined system is further designed.In terms of hardware,each module of the system is selected,the relevant interface circuit is designed,and the PCB board is drawn to complete the construction of the hardware platform;In terms of software,the design of each subroutine is completed in turn according to the functional requirements of the system.Through the actual test experiments,the software and hardware platform designed in this paper is verified.The results show that the IELM-STKF information fusion algorithm designed in this paper can suppress the divergence of system error and effectively improve the positioning accuracy of the system when the Beidou signal is temporarily out of the lock.The integrated navigation and positioning system designed in this paper has certain reference significance for engineering applications.