Satellite attitude acquisition using dual star sensors with a bootstrap filter

We propose a new attitude acquisition method based on the Bayesian bootstrap filtering approach using dual star sensors. The proposed method estimates the right ascension and the declination pointed by each star sensor in the reference coordinate system using the measurement of the number of stars inside the FOV (field of view). The system and the measurement models are highly nonlinear functions of the state. Moreover, measurements are drawn from a finite nonnegative integer set rather than from real numbers, and the well-known extended Kalman filter cannot be used. We propose to apply the Bayesian bootstrap filtering technique assuming that the measurement noise has an arbitrary probability mass function. According to our simulation, the proposed method is computationally faster, and acquires the attitude quickly compared with more traditional triangular point-pattern matching techniques.