Comparison of the extended and unscented Kalman filters for satellite motion states estimation

This paper provides a performance comparison between the extended and unscented Kalman filters for the localization of a spacecraft. The unscented filter uses a well chosen set of sample points to approximate the probability distribution more accurately than the first-order linearization of the extended Kalman filter. Firstly, a brief description and comparison of the two filters is given. Then, a dynamic model of the spacecraft accounting for several perturbing effects is described. The measurement process is based on distance and angle of the satellite with respect to the inertial reference system. Finally, Position and attitude of the satellite are estimated through the above mentioned estimation algorithms implemented in MATLAB. The simulation results show that the unscented Kalman filter is a superior alternative to the extended Kalman filter for the nonlinear satellite motion estimation problems.