Analysis of filtering methods for satellite celestial navigation
地球卫星自主天文导航滤波方法性能分析

The filtering method becomes the most important factor affecting the navigation accuracy and timeconsumption, whenever the accuracy of sensors can not be improved anymore. In satellite navigation, the extended Kalman filter(EKF), unscented Kalman filter(UKF) and unscented particle filter(UPF) are the three widely used filtering methods. The performance of the system based on these three methods under different conditions is analyzed. Hardware-in-loop tests show that in the same condition, the UPF provides the highest navigation performance but requires the most computation; meanwhile the EKF gives the lowest navigation performance but needs the least computation. The conclusions drown by this study are useful in the design and analysis of autonomous navigation system of satellites.