Abstract:

Satellite autonomous orbit estimation is the necessary condition for autonomous existence of the satellite. Observed quantities which are from the commonly used satelliteborne sensors are conversed, so the observation equation is linearization, and the amount of calculation is decreased. Geocentric distance is calculated by using the starlight refraction when the star vector is at a special position, then the amount of information of observation equation is increased. One of the key factors that influence the performance of the filter is that the two body dynamic orbit model of first order recursion approximation is introduced into the system model error. The RungeKutta method is effectively combined into the orbit estimator, and the accuracy is greatly improved. The results of simulation prove that the performance of this high order nonlinear orbit estimation filter is far superior to that of the expansion of the Kalman orbital estimator.