Fault self-checking method of Kalman filter and its application for initial alignment of SINS
Kalman filter is commonly used in initial alignment and integrated navigation of inertial navigation system.Under the condition of abnormal system noise and observation noise,the filter will appear to decline in accuracy or even divergence.In order to solve this problem,two anomaly test criterias for filter based on covariance matching of the new information and Mahalanobis distance chi-square test are analyzed in theory.Considering the typical applications of the proposed criteria in conventional Kalman filter,unscented Kalman filter and cubature Kalman filter,a test is designed based on the initial alignment of strapdown inertial navigation system.Performances of these two test criteria methods under the condition of abnormal system noise and observation noise are verified.However,the strictness of the test criteria is different and the tailing effect have appeared.Therefore,some suggestions in practical application are also given.
inertial navigation systeminitial alignmentKalman filteranomaly test criteria
万方数据
维普
