Study Of Inertial/Gravity-matching Integrated Navigation Algorithm

The inertial/gravity matching integrated navigation is known as a passive navigation method with the advantages of high concealment and high anti-interference ability and hence it is used in many fields like navigation and underwater carriers.Gravity matching algorithm is the core of integrated navigation system.High-precision gravity anomaly map is the precondition for the implementation of gravity matching algorithm.The selections of matching area have an influence on the results of the gravity matching algorithm.Therefore,this paper studies the refinement method of gravity anomaly map,the gravity matching algorithm based on correlation extremum algorithm,and the selection criterion of gravity matching area.The specific research contents are as follows:Three interpolation methods are used to interpolate the gravity map,including inverse distance weighted interpolation(IDW),Kriging interpolation,and radial basis function interpolation(RBF).The interpolation results of different kernel functions are analyzed and compared using the method of cross-validation,which verified the excellent performance of RBF method using the thinplate spline basis function and the cubic surface basis function in the gravity anomaly map interpolation.The RBF based on the cubic surface basis function is selected as the best to refine the gravity anomaly map,which paves the way for the research of the gravity matching algorithm.The difference method is used to reduce the influence of the Etovos effect and other systematic errors.A two-step correlation extremum gravity matching algorithm is proposed,which combines the rough matching and appropriate matching using rigid transformation.The effect of this is verified through the simulation experiments: The accurate search method of INS trajectory is not suitable for large-scale search range,and the fast MSD method may have a large number of matching points with excessive error.The two-step matching combines the merits of the two methods and is adapted to the case where the initial error of the INS is large.The efficiency and real-time performance are improved,while the high-precision matching result is maintained.The statistical analysis of the characteristic parameters of the gravity field is carried out.This paper compares the empirical selection criterion,the analytic hierarchy process(AHP),the entropy method,the grey relational analysis method,and principal components analysis(PCA).The weights of the AHP and entropy are introduced into the gray relational analysis,and a combination of the analytic hierarchy process and the entropy method is proposed.The effectiveness of these criteria is validated by correlation analysis and experiments.