Performance Comparison of GNSS/INS Integrations Based on EKF and Factor Graph Optimization

Abstract

Integration of global navigation satellite system (GNSS) and inertial navigation system (INS) is extensively studied in the past decades. Conventionally, the two most common integration solutions are the loosely and the tightly coupled integrations using extended Kalman filter (EKF). Recently, the factor graph technique is adopted to integrate the GNSS/INS and improved performance is obtained compared with the EKF-based GNSS/INS integration. However, only simulated data are tested to show the effectiveness of factor graph-based method in the existing work. Moreover, the reason that why the factor graph-based integration obtains better performance is not presented in the existing reference. Therefore, this paper proposes to compare the performance of EKF, and the factor graph-based GNSS/INS integrations. Both loosely and tightly coupled integrations are comprehensively discussed. We test the four different GNSS/INS integration methods in typical urban scenario in Hong Kong. The performances of the four solutions are compared. The conclusion shows that the factor graph-based tightly coupled GNSS/INS integration obtains the best performance among the four methods. The detailed analysis of the reasons for the improvement caused by factor graph is also given in the paper from the angles of re-linearization and iteration.