Distributed Radar Multiframe Detection With Local Censored Observations

Distributed radar multiframe detection (DR-MFD) provides an efficient approach for detecting dim targets. However, the transmission and processing of a large number of measurements in the DR-MFD lead to huge computation and communication costs. To mitigate the abovementioned burdens, this article investigates the DR-MFD problem with local censored observations, and at each frame, only local plots with strength over a threshold are transmitted to a fusion center. The measurement errors and asynchronous searching of local radars make the received plots misaligned, then it is difficult to fuse the plots before data registration that is usually neglected in existing studies. To tackle this difficulty, a test window-based method is proposed for data registration, and then, a generalized likelihood ratio detector is derived to fuse the multiradar multiframe plots. Simulation results indicate that the proposed method can efficiently handle the data registration problem with a low computational cost. It is also shown that for most cases, the signal-to-noise ratio loss caused by the censoring procedure is less than 1 dB while the communication cost is reduced to 10%.