Adaptive Robust Fading Memory Filter for Tracking Maneuvering Spacecraft with Heavy-Tailed Measurement Noise

This paper proposes an adaptive robust fading memory filter (ARFMKF) to accurately track impulsively maneuvering spacecraft under heavy-tailed measurement noise. The ARFMKF is capable of simultaneously suppressing the adverse effects of measurement outliers, heavy-tailed measurement noise, and impulsive maneuvers. First, the Huber robust M-estimation methodology is investigated to mitigate the influence of measurement outliers and heavy-tailed measurement noise, enhancing the filter’s robustness to measurement model uncertainties. Then, an adaptive fading factor is incorporated to improve the maneuver detection sensitivity of the proposed filter and is used as a weighting coefficient for covariance inflation to compensate for the dynamics model errors caused by impulsive maneuvers. In addition, a discrimination index is designed to effectively distinguish impulsive maneuvers from measurement outliers, ensuring appropriate covariance inflation in the presence of maneuvers. Simulations show that the proposed ARFMKF outperforms the comparison filters in terms of maneuver detection sensitivity, discrimination capability, and overall tracking performance, particularly in scenarios involving multiple impulsive maneuvers and frequent measurement outliers.