Application of the EM algorithm for the multitarget/multisensor tracking problem

Abstract

An important problem in surveillance and reconnaissance systems is the tracking of multiple moving targets in cluttered noise environments using outputs from a number of sensors possessing wide variations in individual characteristics and accuracies. A number of approaches have been proposed for this multitarget/multisensor tracking problem ranging from reasonably simple, though ad-hoc, schemes to fairly complex, but theoretically optimum, approaches. In this paper we describe a new iterative procedure for multitarget/multisensor tracking based upon use of the expectation-maximization (EM) algorithm. More specifically, we pose the multitarget/multisensor tracking problem as an incomplete data problem with the observable sensor outputs representing the incomplete data while the target-associated sensor outputs constitute the complete data. This formulation then allows a straightforward application of the EM algorithm which provides an iterative solution to the simultaneous maximum-likelihood (ML) and/or maximum a posteriori (MAP) estimate of the target states, under the assumption of appropriate motion models, based upon the outputs of disparate sensors. The advantage of this EM-based approach is that it provides a computationally efficient means for approaching the performance offered by theoretically optimum, but computationally infeasible, simultaneous ML estimator. We provide selected results illustrating the performance/complexity characteristics of this EM-based approach compared to competing schemes.This work supported in part by DAAH04-95-1-0103

Recovery of the trajectories of multiple moving objects in an image sequence with a PMHT approach

“…[17], the authors propose a recursive scheme closely related to PMHT in which the association variables form a Markov random field. The method we have designed remains, as in Ref.…”

An EM‐based adaptive multiple target tracking filter

“…A parameter that controls the complete data is the state (k) as used by Molnar [23]. In addition, we de"ne a parameter (k) that includes the various sources of external in#uence.…”

A constrained optimal data association for multiple target tracking

“…Among many h4l'T schemes, the three most important methods are Joint Probabilistic Data Association (JPDA) [l], the Expectation Maximization approach [9], and the neural net approach [lo, 61. JPDA is a kind of closed loop system consisting of two systems: data association and prediction. For each time frame, Kalman filters predict target centers.…”

“…After that, the updated measurements are used in the Kalman filters again. Molnar [9] derived the whole system consisting of prediction and association units in an unified manner with the EM method [3]. Defining the association matrix as missing data, the algorithm can estimate the association matrix for the current measurements.…”