Failure Detection Method Based on Fuzzy Comprehensive Evaluation for Integrated Navigation System

摘要

A failure detection method based on fuzzy comprehensive evaluation for integrated navigation systems was presented in this paper. By using fuzzy comprehensive evaluation, this method judged the measured data of every subsystem of a Kalman filter comprehensively in order to determine their states. This method overcame some shortcomings of conventional failure detection methods. At last, the contrastive experiment among this method, chi-square test and the method of data change rate has indicated that chi-square test could recognize the larger outliers, but couldn't detect the lesser, that the method of data change rate couldn't deal with the longtime continuous failures, and that the method based on fuzzy comprehensive evaluation could recognize all of the failures. So it was a quite practical method.