Robust detection of distributed CA-CFAR in presence of extraneous tagets and non-Gaussian clutter

Abstract

This paper deals with distributed CA-CFAR detection in presence of Gaussi an and non-Gaussian clutter. In Gaussian environment, we propose to apply a wavelet transform based on soft-thresholding in multisensor CA-CFAR systems employing parallel decision fu­ sion in both homogeneous and non homogeneous background in the sense of the Neyman-Pearson (N-P) test. In that context, we propose two approaches to combine the data from the different CA-CFAR detectors to achieve even better detection performance. In the non-Gaussian environment, we propose another preprocess­ ing approach, based on a non-linear compressing filter, to reduce the noise effect. The three proposed new methods are shown to provide better detection performance, especially in lower SNR and in the presence of extraneous targets and heavy-tailed noise_