Random Matrix based Extended object/Group Target Tracking with Extra Extension Parameters

The extended object (EO) and group target (GT) tracking based on random matrix can estimate the kinematic state and extension jointly. In order to describe the more complicated deformation of extension, in this paper, we add extra extension parameters to the extension state and construct a generalized random matrix model. To estimate the joint state consisting of kinematic state, extension, and other extension attributes, a sequential Monte Carlo technique called Rao-Blackwellized particle filter (RBPF) is introduced to approximate the posterior density of joint state due to the complexity of the new model. Simulation results show that the proposed model and the new approach have obvious improvement in estimation performance compared with the previous approaches.