Performance Analysis of eXogenous Kalman Filter for INS/GNSS Navigation Solutions

There are several methods of fusing data for navigation solutions using Inertial Navigation System (INS) aided by Global Navigation Satellite System (GNSS). The most used solutions are nonlinear observer (NLO) and extended Kalman filter (EKF) of various architectures. EKF based estimation methods guarantees sub-optimal solutions but not stability, on the contrary NLO based estimation guarantees stability but not optimality. These complimentary features of EKF and NLO has been combined to design an eXogenous Kalman filter (XKF) where the estimate from the NLO is used as an exogenous signal to calculate the linearized model of the EKF. The performance of the designed XKF is tested on real flight test data collected using a Slingsby T67C ultra-light aircraft. The results show that during the outage of GNSS, in some cases the divergence of position estimates using XKF is lower compared to EKF and NLO, however no clear benefit is achieved.