Autonomous Navigation and Observability Analysis for Satellites During Orbital Maneuver

One of the basic requirements of the satellite’ s autonomous operation is autonomous navigation, which aims at determining the position and velocity of the satellite according to those onboard sensors only. However, during the orbital maneuver, designing an optimal estimator is difficult because the thrust acceleration of the satellite is not precisely known. Based on the observability theory of nonlinear systems, the observability of autonomous navigation systems with unknown thrust acceleration was proved. The simulation results show that when the thrust is modeled as a constant vector, the augmented EKF algorithm can achieve accurate estimation, while it does not perform well when the thrust is modeled as a time-varying vector. An estimation algorithm based on a neural network observer was designed to solve this problem. Simulation results show that the proposed algorithm, which only uses position measurements, can accurately estimate the satellite motion during an orbital maneuver with time-varying thrust acceleration.