Distributed navigation information fusion method based on information geometry

Abstract

Due to the different inherent defects of various single navigation systems, the application of information fusion methods to synthesize information from multiple navigation sources has become a hot issue in the field of navigation and positioning. In this paper, a multi-navigation source fusion method based on information geometry (Kullback-Leibler Divergence Minimization, KLM) is proposed, which maps the information accuracy probability function of each navigation source to the Riemann space, and establishes the navigation source information probability set manifold. The optimal fusion of the information geometric manifold of the navigation source under the Riemann information geometry architecture. This method uses the K-L divergence to replace the geodesic distance, which greatly reduces the amount of calculation and effectively improves the overall positioning accuracy. The simulation results show that after the information of each navigation source is fused by the KLM method, the positioning error obtained after fusion is significantly reduced. And the more fusion navigation sources, the higher the positioning accuracy. In the 3D multi-source fusion positioning scenario, the positioning accuracy is improved by more than 30% on average.