| With the increasing demands for navigation accuracy,multi-source navigation has gradually become the preferred method for space missions.Landmark navigation came into being with the development of optical technology,and compensates with CNS to improve the navigation accuracy of integrated navigation.At the same time,with the development of intelligence algorithms,the navigation filter algorithm becomes more adaptive.Therefore,multi-source and adaptability have become the new directions of navigation,and high stability has always been the goal of navigation.This paper uses multi-source navigation as the background,and the work done by this paper is as follows:(1)Firstly,in the preprocessing section,the navigation data is detected and compensated by BP neural network,and the processed information with continuity and obvious characteristics is obtained.(2)Taking the landmark navigation as the research point,and in view of the position as well as attitude determination of CNS,a navigation method using the landmark camera is proposed.In the simulation section,compared with the results that without attitude determination,it is found that this method is more effective,because of the image information of the camera is made full use of.(3)In the research of multi-source navigation,the federated filter is used as the structure,and Sage-Husa adaptive filter is used to improve the federated filter to obtain a federated adaptive filter,which satisfies the characteristics of adaptive and distributed algorithm.In the simulation part,the improved federated adaptive filter is compared with two separate methods,and the improved algorithm has more obvious advantages;(4)Finally,in the SINS/GNSS/CNS/landmark integrated navigation scheme,the BP neural network is used to improve the navigation results for the uncertainty problem in multi-source navigation.The simulation results show that the improved method using BP neural network is more accurate. |
PDF Full Text Request