Robust Extended Kalman Filter for GPS/INS Relative Navigation

In this paper, a robust extend Kalman filter was proposed for GPS/INS integrated relative navigation systems between cooperative target and tracker in the presence of inertia uncertainties. Firstly, the first-order of Taylor series expansion was utilized to linearize the nonlinear term of the relative navigation systems approximately, while the model error caused by linearization was considered as the uncertainties of the system. Then, based on the robust Kalman filter, a robust extend Kalman filter algorithm was proposed for the GPS/INS integrated relative navigation system. Finally, the simulation result indicates that the proposed method provides relative position accuracy of 0.1 m and relative attitude accuracy of 0.001°, which manifests a high accuracy and a strong robustness against the inertia uncertainty of tracker.