Autonomous optical navigation based on Adaptive SR-UKF for deep space probes
Lichao Ma, Liu Zao-zhen, Xiuyun Meng · Chinese Control Conference · 2010
A novel adaptive Square-Root Unscented Kalman Filter (Adaptive SR-UKF) was proposed to solve the problem that the priori noise statistics are difficult to obtain accurately in an autonomous optical navigation for deep space probes. The adaptive SR-UKF realizes its initialization by using an approximation of the priori noise statistics, which is corrected at each step by the adaptive SR-UKF that adopts a limited memory algorithm for the estimation of the noise statistics based on noise samples. Effectiveness of the Adaptive SR-UKF is examined by the simulation of an autonomous optical navigation for the cruise phase of a cislunar probe, and results show that the Adaptive SR-UKF is much superior to UKF and SR-UKF when the priori noise statistics are inaccurate.