Optimization analyses of Kalman and H∞ filters in SINS/GPS integrated navigation systems

摘要

For this paper, the colored noise conditions were analyzed for optimal estimation of system error for integrated navigation systems by using the Kalman and Hoo filtering algorithms. The integrated navigation system uses the resulting difference between Strap-down Inertial Navigation (SINS) and Global Positioning System data (GPS) as the input value for the filter. The errors in the integrated navigation system are estimated and corrected by using one of the filtering methods in real time. The corrected SINS data output for the integrated navigation system is used to verily the effectiveness of the two kinds of filtering algorithms. The results showed that Kalman filter's velocity and position estimate error are larger than that of the H∞ filter, with Kalman filter's velocity estimates error range being up to 0.9m/s, while the H∞ filter's maximum error is only 0.3m/s. Thus, the H∞ filtering algorithm has better stability and robustness.