A survey on joint tracking using expectation–maximization based techniques

Abstract

Many target tracking problems can actually be cast as joint tracking problems where the underlying target state may only be observed via the relationship with a latent variable. In the presence of uncertainties in both observations and latent variable, which encapsulates the target tracking into a variational problem, the expectation-maximization (EM) method provides an iterative procedure under Bayesian inference framework to estimate the state of target in the process which minimises the latent variable uncertainty. In this paper, we treat the joint tracking problem using a united framework under the EM method and provide a comprehensive overview of various EM approaches in joint tracking context from their necessity, benefits, and challenging viewpoints. Some examples on the EM application idea are presented. In addition, future research directions and open issues for using EM method in the joint tracking are given.

Variational Probabilistic Multi-Hypothesis Tracking

“…Inspired by the literature mentioned above, this paper aims to develop an MTT algorithm capable of handling track-loss while maintaining the strong points of PMHT. It is believed that the limited track-loss performance of PMHT is led by the expectation-maximisation (EM) algorithm applied, which requires the exact number of mixture components (number of targets in MTT) as prior knowledge [32,33]. In variational Bayesian expectationmaximisation (VBEM), the number of mixture components can be set large, and through optimisation, the components that provide insufficient contribution describing the dataset will have their weight converging to zero [34].…”

Joint Target Detection and Tracking in Multipath Environment: A Variational Bayesian Approach

“…The EMbased algorithms in [20]- [22] alternate between computing the expected complete log-likelihood according to the posterior probability density function (PDF) of missing data (multipath data association) in E-Step and optimizing it with respect to (w.r.t.) the model parameters (state estimation) in M-Step, which are attractive and desirable to reduce the performance deterioration caused by the coupling between identification errors (from data association) and estimation errors [23]. However, these work considered a single target tracking.…”