Censoring in distributed radar tracking systems with various feedback models

In this paper, various methods to control censoring of local state estimates in a distributed multisensor radar tracking system are proposed. The data flow architectures that are used include three different feedback methods in order to achieve the goal of adequate state estimation using the fewest number of local state estimates sent to the fusion center as possible. The main novelty introduced in this paper is the use of J-Divergence for censoring the local state estimates in conjunction with the various feedback architectures. A simulation was run on these architectures with three spatially diverse sensors and a simple non-maneuvering target. Results show that distributed architectures that provide feedback from the fusion center to the local censoring processes or to the local state estimators, yield better estimates with more one-way communication savings.