Improved Interacting Multiple Model Expansion Box Particle Bernoulli Filter Algorithm

For the problem of maneuvering target tracking in clutter environment, an improved Interacting Multiple Model Expansion Box Bernoulli filter (IMM-Box-BF) is proposed by combining interacting multiple model (IMM) with Box-Bernoulli filter (Box-BF). The predicted box particles are divided into small box particles, and then expanded into a certain interval length. The simulation results show that the proposed improved IMM-Box-BF has more accurate state estimation and more stable tracking target than IMM-Box-BF and Partitioned IMM-Box-BF.