Multi-target tracking in a two-tier hierarchical architecture

In this paper, a two-tier hierarchical architecture is proposed to address the multi-target tracking problem using a particle probability hypothesis density filtering algorithm. According to a proposed cluster scheduling method, the base station selects active clusters at each time step and determines their order for the sequential data fusion in the second level of hierarchy. Within each active cluster, sensors transmit their measurement-sets to the cluster head, which processes the information locally and estimates the number of targets and their states. The proposed architecture works well even when the target dynamics and/or measurement process is severely nonlinear. The performance of this architecture is demonstrated in the application of bearing and signal strength tracking.