Autonomous deep space navigation based on fading memory filtering

A new navigation Scenario had been designed and analyzed to autonomously determinate the orbits of two spacecrafts when they are circling the moon, based on measurements of the relative positions vector from one spacecraft to the other. In this article, more realistic and complex orbital perturbation factors had been considered, which would cause a divergent result due to the errors in the model of dynamics and process noises. Accordingly, a reformative square-root Uncented Kalman filtering (SRUKF) based on fading memory method was proposed, which was able to resolve this divergent problem in a good condition. Viewed from the simulation results, the new estimated arithmetic performs well as that was expected.