Information Fusion Progress: Joint Optimization Based on Variational Bayesian Theory
Quan Pan, Yumei Hu, Hua Lan, Shuai Sun, Wang Zeng-fu, Yang Feng · RMIT Research Repository (RMIT University Library) · 2019
By reviewing the development of information fusion theory in recent years, this paper analyzes the problems of target tracking systems, such as nonlinearity, multi-mode, deep coupling, networking, high-dimensionality and unknown disturbance input, and points out the necessity of joint optimization in target tracking system. Furthermore, several joint optimization methods, including the joint detection and estimation, joint clustering and estimation, joint association and estimation, joint decision and estimation are discussed. Meanwhile, we emphatically introduce the integrated optimization method based on the variational Bayesian theory that provides a unified framework of joint identification and estimation. Taking over-the-horizon radar as an application background, we give a general joint optimization method for the multi-path multi-mode multi-target tracking system in this paper. In addition, future research directions of the variational Bayesian theory in the field of target tracking are discussed.