Variational Approximation for Extended Target Tracking in Clutter with Random Matrix
Xiaojun Yang, Qinqin Jiao · 2021 International Conference on Control, Automation and Information Sciences (ICCAIS) · 2021
Random matrix is a simple and efficient method for extended target tracking. In this paper, based on variational Bayesian approximation, we present a Gamma Gaussian inverse Wishart filter for the extended target tracking in clutter with unknown measurement rate. A new probability data association method based on random matrices is proposed to handle the measurement origin uncertainty. The recursive estimation of the kinematic state, extension and unknown parameters are derived using variational Bayesian inference. Finally, the simulation results verify the validity of the proposed method.