Research On Low Cost Vehicle Integrated Navigation System Based On Multi-sensor Information Fusion

When low-cost MEMS based inertial measurement unit(MIMU)is used for inertial navigation system(INS),the error will accumulate quickly because of the large MEMS sensor measurement noises.Therefore,its combination with other navigation methods has become the main research direction of land vehicles'navigation systems.At the same time,the recent advances in sensor fusion technology contributed to the realization of the highly integrated low-cost navigation systems.This paper focuses on the key technologies of low cost MIMU/WSS/MAG integrated system.The main contents of this paper are as follow:1.The Strapdown Inertial Navigation System(SINS)attitude,velocity and position updating algorithm are deduced,considering the cone motion error and rowing error compensation in the process of attitude and 'velocity updating,respectively.Finally,using the measurement of MIMU in static conditions,the pure inertial navigation results show that when using MIMU for INS,the pure inertial navigation system has good accuracy in the first 10s,but the posture,velocity and position error will increase rapidly later.2.In order to solve the initial alignment problem of low-cost integrated navigation system,the magnetometer(MAG)aided initial alignment algorithm is studied in this paper.First of all,for the complex vehicle environment,this paper analyzes the soft iron and hard iron effect of the magnetometer,and uses the ellipsoid fitting algorithm to compensate the interference caused by the magnetometer.By comparing the root mean square values of the output of the magnetometer before and after the compensation,the RMS value of the output of the compensated three-axis magnetometer is close to 1,which proves that the azimuth error of the magnetometer is well compensated.And then the accelerometer and magnetometer are used to achieve a rough alignment,providing more accurate initial attitude angle for fine alignment and greatly shorten the initial alignment time.Finally,the initial precision algorithm based on the traditional Kalman filter theory is designed.The initial alignment experiments under the static base show that the proposed initial alignment algorithm can converge in a short time.A MIMU/WSS/MAG integrated navigation system is designed.This paper mainly describes the design process of the federated Kalman filter because the FKF has advantages of less computation and centralized fault detection,diagnosis and isolation for the navigation subsystem.SINS/WSS and MIMU/MAG subsystems are designed separately after careful analysis of the navigation subsystem combination mode.The state equations and observation equations of each subsystem are established.Then,the two sub-systems are carried out according to the principle of information distribution of federal Kalman filter.Finally,the actual vehicle experiment is designed.By comparing the results of federated Kalman filter and centralized Kalman filter,the advantages of federated Kalman filter in navigation accuracy and fault detection and isolation are verified.This paper provides a systematic study on the low-cost MIMU/WSS/MAG vehicle integrated navigation system,which provides a feasible combination scheme for low-cost vehicle-mounted autonomous navigation.It provides a new method for further realization of low-cost and highly integrated vehicle integrated navigation system.