Satellite Tracking Using the Space-Based Optical Sensor and Shifted Rayleigh Filter

Space-based optical sensors play an increasingly important role in space surveillance. Due to the limited computation resource on a satellite, an accurate and efficient onboard target tracking algorithm is desirable. Traditionally, the extended Kalman filter (EKF) is preferred because of its low complexity. However, the EKF may become divergent when the initial target state error is large. In this paper, a novel satellite tracking algorithm is proposed. It adopts the linearized satellite dynamics to propagate the state prior mean and covariance, and adopts the shifted Rayleigh filter (SRF) to perform the data assimilation. The SRF changes the conventional measurement model to a displacement vector plus noise form. By carefully choosing the noise covariance, the measurement probability density functions of this new and the traditional models are matched with each other. A numerical simulation is conducted and the results show that the proposed algorithm has superiorities in tracking accuracy and filter consistency compared with the EKF, at the cost of slightly increased computation time. Thus, the proposed algorithm is of practical value.