Orbit and attitude state estimation accuracy of an autonomous satellite navigation system

Monte Carlo analysis is employed to evaluate the ability of an autonomous satellite navigation system to determine its orbital and attitude states. Estimation of the orbit and attitude states are treated in a coupled fashion. The navigational system considered consists of a sun sensor, a landmark sensor, and a three-axis rate gyro package. The observation data are processed in a batch-sequential filter which decomposes into an extended Kalman filter as the duration of the batch interval becomes less than the time between two consecutive observations. To avoid singularities in the propagation of the attitude states, quaternions were used to represent the rotation from the body-fixed axes to the inertial coordinate system. However, the attitude was estimated in terms of Euler angles to avoid a singular covariance matrix. Test cases indicate that the estimates of the position in the transverse direction contain the greatest uncertainty