Performance of decentralized L-CFAR detection in inhomogeneous background

The problem of decentralized constant false alarm rate (CFAR) detection, with L-CFAR remote sensors, is considered. Following a previous paper, concerning the design and the analysis of such detection structure in homogeneous environment, we now assume that a detector, designed to operate with homogeneous background, is affected instead by clutter edges and secondary targets entering the local sensor reference windows. For this inhomogeneous environment, we derive closed-form analytical approximations of the detector performances. Results of computer simulations, which validate the derived formulas, are provided.