Research of interacting multiple model particle filter based on passive multi-sonar

The passive target tracking problem of multi-sonar belongs to nonlinear problem, and the background noise is not guass noise entirely, so a kind of filtering method which suit for nonlinear and non-guassian system of is needed. In this paper, the interacting multiple model particle filter which combined by particle filter and interacting multiple model is applied in tracking maneuvering target. Combing the particle filter´s ability of disposing nonlinear and non-guass problem and the IMM´s ability of tracking target´s arbitrary maneuvering, the difficulty that traditional filter encounters in tracking maneuvering target under the circumstance of nonlinear and non-guassian can be solved. The interacting multiple model particle filter is simulated in two different model set, the results indicate that, under the case of model matching, this method has good ability in tracking the maneuvering target of nonlinear and non-guassian system.