Application of fuzzy adaptive filtering algorithm in the accuracy evaluation of transfer alignment

Abstract

During the process of accuracy evaluation of transfer alignment, since the parameters of the Inertial Navigation System (INS) model usually have some deviations from that of the actual physical process, and the noise statistic characteristic is not exactly known and the equipment environment in the evaluation process affects the results, the normal Kalman filter may behave badly or diverge and we get the results which are not accurate as we expected. For the purpose of estimating accuracy and stability of transfer alignment and solving divergence problem caused by signal loss during the transfer alignment process, a fuzzy adaptive filter with fixed-point smoothing and fixed-interval smoothing algorithm was proposed and it might inhibit model divergence, improve the evaluation accuracy and reduce the amount of calculation. Finally, the experimental simulation showed that, compared with the fixed-point smoothing and fixed-interval smoothing algorithm based on typical Kalman filtering algorithm, the fuzzy adaptive filter not only strengthened the filtering convergence capability, but also improved the evaluation accuracy, which could efficiently evaluate the accuracy of transfer alignment.