Redundant adaptive robust tracking of active satellite and error evaluation

An approach to active target satellite tracking is presented. The kinematic models based on differential orbital elements and unbiased converted measurements are adopted to approximate the dynamic orbital relative tracking system, and then followed by the proposed redundant adaptive robust extended Kalman filter (RAREKF). The analysis of working status of robust filters shows that the innovation controlled switching scheme in AREKF fails under the unmodelled non-linear errors that exist all the time in orbital motions, and the loss of filtering optimality will be caused. So, a tunable redundancy factor and a novel compensation function are introduced into the RAREKF by extending the sufficient conditions of stable filtering. To evaluate the result of this tracking method, a meaningful error index is provided by consideration of the errors from both the model and the filter. The simulation shows the superiority of the RAREKF over the AREKF, uncoupled extended Kalman filter (EKF) and the usual EKF. By using the provided index, it also illustrates the lowest error level and best performance of the proposed tracking method in all the compared ones with different models and filters.