Distributed multi-view multi-target tracking based on CPHD filtering

This paper addresses distributed multi-target tracking over a network of sensors having different fields-of-view (FoVs). Specifically, cardinalized probability hypothesis density (CPHD) filter is run at each sensor node. Due to the fact that node has limited FoV, standard fusion methods need be suitably modified. In fact, monitored area multiple nodes consists several parts are either exclusive single node, i.e. or common (at least two) nodes. this setting, crucial issue how account for these information sets in rule. The problem particularly challenging when knowledge FoVs unreliable, example because presence unknown occlusions, and local can have valuable also on targets located outside current nominal thanks diffusion time-varying. context, we propose CPHD based idea decomposing posterior densities fused so as sets. A new decomposition method proposed does not rely and, instead, uses clustering decompose intensity function into sub-intensities, reconstructing corresponding cardinality distribution via multi-Bernoulli approximation. Then, performed parallel according geometric average arithmetic Simulation experiments provided demonstrate effectiveness approach.