A novel adaptive square root recursive least squares filter algorithm of SINS/SAR integrated navigation systems

In strapdown inertial navigation system (SINS) synthetic aperture radar (SAR) integrated navigation systems, the time-varying parameter estimation is a challenging problem because SAR measurements are fairly few and have unequal interval. In such case, the common Kalman filter algorithm can´t be convergent. Furthermore, the traditional recursive least squares (RLS) algorithm is difficult to estimate the time-varying parameter. This paper proposed a novel adaptive square root recursive least squares (ASRRLS) filter algorithm of SINS/SAR integrated navigation systems. The first characteristic of ASRRLS is introducing adaptive fading factors based on the orthogonality principle, which can effectively track the parameters´ varying. The second characteristic of ASRRLS is taking covariance square root matrix instead of covariance one in the filter algorithm, which can avoid filtering divergence and improve the filter convergence speed. The ASRRLS filter algorithm is deduced and the performance is rigorously analyzed. Simulation results demonstrate the feasibility and effectiveness of the proposed approach.