Target Tracking for Maneuvering Objects Using Kalman Filter with Interacting Multiple Model (IMM) Algorithm

This paper discusses tracking methods for maneuvering targets exhibiting either uniform linear motion or uniform circular motion. To address this problem, we establish target motion models and observation models. The core implementation utilizes a Kalman filter integrated with the Interacting Multiple Model (IMM) algorithm, which operates through parallel filters for different motion modes with interactive probability weighting. The algorithm structure involves model-set design (e.g., constant velocity and coordinated turn models), model transition probabilities, and mixing of state estimates. Simulation experiments confirm that the algorithm not only accurately tracks both uniform linear and circular motions but also maintains minimal filtering errors during motion model transitions. The IMM framework demonstrates particular effectiveness through its recursive model probability updates and soft-switching mechanism between filters. Thus, this approach proves highly effective for handling maneuvering target tracking challenges. Keywords: Kalman filter; target tracking; maneuvering; Interacting Multiple Model (IMM)