A Method for Lever Arm Estimation in INS/GPS Integration Using Direct Unscented Kalman Filter

Abstract

In Inertial Navigation System (INS) and Global Positioning System (GPS) integrated system, lever arm effect is an important error source, which makes the integrated system highly nonlinear. In addition to nonlinear problem, the lever arm is difficult to measure in practical applications. To solve the problems, the direct filtering method considering the lever arm effect of Unscented Kalman Filter (UKF) is proposed in this paper. The proposed method adds lever arm to system model of direct Kalman filter and compensates the estimated lever arm, and thus the lever arm is constantly estimated and modified as the filter estimations are updated. Specifically, the system state model and measurement model are established based on the direct filtering method. The navigation parameters of INS and lever arm are taken as system state variable, the velocity and position of GPS are taken as measurement variable. Then the UKF is used for INS/GPS integrated system information fusion. Simulation results show that the proposed direct UKF considering lever arm of integrated navigation system can estimate the lever arm correctly. Furthermore, the accuracy of the proposed method is significantly higher than that of standard indirect KF and standard direct UKF.