Abstract:

 In this paper, a modified quasi-consistent cubature Kalman filter (MCKF) based on relative position measurements in a dual-satellite system is proposed for the autonomous navigation of a distributed satellite system (DSS). The only difference between the MCKF and the standard cubature Kalman filter (CKF) is that in MCKF the error covariance matrix from the observability theory of a nonlinear system is adjusted on line. The numerical simulations show that the MCKF is quasi-consistent and less sensitive to the observable degree than the usual one. Moreover, it indicats that the MCKF can shorten the stabilizing time of the systems with high accuracy.

Key words:  Autonomous navigation, Cubature Kalman filter (CKF), Observability analysis