Multi-source Fusion Navigation Method Based on Optimized Pre-integrated Factor Graph
To address the issue that traditional factor graph algorithm cannot give fully play to the advantages of high-precision inertial devices,and lead to significant navigation errors,a factor graph multi-source fusion navigation method based on optimized pre-integration is proposed.The inaccuracy of IMU factor modeling in traditional factor graph algorithm is explored,considering the impact of Earth's rotation and the angular velocity caused by carrier motion on algorithm accu-racy.The process of inertial pre-integration traditional factor graph is optimized,the trajectory simulation based on the opti-mized pre-integration method is performed and the positioning accuracy is tested under conditions of good GNSS signals and phased GNSS-denied.The simulation results show that for navigation-grade inertial navigation systems,the root mean square error in the horizontal direction of the factor graph algorithm based on optimized pre-integration is less than 1 m.Un-der the condition of good GNSS signal,the horizontal positioning accuracy of the factor graph algorithm based on optimized pre-integration is improved by 8.57%and 9.52%compared with the EKF algorithm and the traditional factor graph meth-od,respectively.In case of phased GNSS denial,the horizontal positioning accuracy is improved by 28.29%compared with traditional factor graph method.And the horizontal positioning accuracy is improved by 24.83%compared with the traditional factor graph method in case of long-term GNSS denial.
NETL
NSTL
万方数据
