Distributed Multiple Resolvable Group Target Tracking Under Limited Field of View Sensors

This paper considers the challenging problem of tracking multiple resolvable group targets (RGT) using sensors with limited field of views (FoVs). To perform RGT tracking in distributed multi-sensor networks, firstly, we run a group structure augmented Labeled Multi-Bernoulli filter (GSA-LMB) on each sensor for accurate estimates of target states and group structures. Secondly, we propose a novel fusion algorithm for RGT tracking which not only fuses local multi-target state estimates but also considers local group structure estimates provided by the GSA-LMB filter. Specifically, the local group structure estimates are incorporated into the track consensus algorithm via the graph matching method, thereby enhancing the accuracy of track matching between two nodes. Lastly, numerical simulation and experimental results demonstrate the effectiveness of the proposed RGT fusion algorithm.