This paper presents several methods to improve the estimation of the star positions in a star tracker, using a Kalman Filter. The accuracy with which the star positions can be estimated greatly influences the accuracy of the star tracker attitude estimate. In this paper, a Kalman Filter with low computational complexity, that can be used to estimate the star positions based on star tracker centroiding data and gyroscope data is discussed. The performance of this Kalman Filter can be increased by adjusting its parameters using certain available information. Using information such as the power in the star signal or the shape of the signal, the noise values in the filter can be adjusted to improve the star position estimate. Furthermore, the filter also estimates the uncertainty on the star positions. These uncertainties can be used to assign more importance to stars with lower position uncertainty in the cost function of the attitude estimation algorithm. The Kalman Filter with these improvements was implemented and tested with simulated star data. Results show that the attitude estimation error is reduced significantly. This results in a more accurate attitude determination and control system which allows to perform more demanding missions.
展开▼
