Tracking of Groups of Targets Using Generalized Janossy Measure Density Function

This paper describes a new approach to tracking multiple groups of targets using the concept of the generalized Janossy measure density function. Each group is modeled by a group state that consists of a group common state and an unknown number of individual target component states. In order to represent a probability distribution of such a group state, we propose to use a generalized version of the Janossy measure density function. We will formulate our tracking problem as a distributed estimation problem in which each local sensor processing unit (node) produces single-frame-based group-level detections and a central fusion center (node) fuses those detections into group tracks. A simple illustrative example will be given