Distributed Information Fusion Particle Filter
Lin Mao, Da Wei Yang · 2014
The method of particle filters as a main solution for non-linear is widely used in digital communication, target tracking, automatic control and signal processing region. In order to eliminate the existing problems such as low precision and low stability, the information fusion is introduced to fuse multiple different sensor measurement information according to a certain fusion criterions. This schematic increases not only the measurement information deterministic and stability, but also the precision and reliability of the particle filter without adding any measurement base stations. The paper proposes an information fusion particle filter algorithm that takes the local particle filter results into distribution fusion utilizing the three weighted information fusion criterions including matrix, scalar and vector (diagonal matrix) methods based on linear minimum variance. Then, a three-sensor bearings-only passive location example illustrates the effectiveness of this proposed algorithm.