The Two-stage Particle Filter for Underwater Small Target Fusion Tracking in Multistatic System

摘要

A two-stage particle filter algorithm is proposed to deal with underwater small target fusion tracking in multistatic system. The equivalent measurements are acquired by the density-based fusion process based on particles sampled by observations from receiver nodes through the first stage particle filter. To tackle the particle impoverishment and improve tracking performance, in the second stage the mixture particle filter implements target state estimation through combining samples from the system observations formed in the first stage and the prediction of last state estimation. The algorithm was utilized for the realistic experiment data processing for the underwater small targets tracking in multistatic sonar system. The proposed algorithm yields more precise and stable state estimation. The processing results show effectiveness and improvement of the two-stage particle algorithm for the underwater small target fusion tracking in multistatic system.