Robust rendezvous navigation in elliptical orbit

The application of a robust state estimation method to the problem of relative navigation during spacecraft in elliptical orbits is discussed. The robust filter is a recursive form of the Huber estimator, which is blend of the least square estimation and the least absolute value estimation techniques. The robust method was compared with the Kalman filter or recursive least square method, using a perturbed Gaussian distribution as the basis for generating random measurement errors. The results demonstrate that the Huber filter outperforms the Kalman filter for non-Gaussian measurement error probability distribution in terms of the state estimation error variances.