Optimization of Distributed CFAR Detection using Grey Wolf Algorithm

In this paper, decentralized constant false alarm rate (CFAR) detection in homogeneous and heterogeneous Gaussian clutter using Grey Wolf Optimization technique is investigated. For independent signals with known power, optimal thresholds of local Greatest Of-CFAR and Smallest Of-CFAR detectors are optimized simultaneously according to a preselected fusion rule. Based on the Neyman-Pearson type test, both the Biogeography Based Optimization and the Grey Wolf Optimization tools are used to conduct distributed CFAR detection comparisons. In terms of achieving fixed probabilities of false alarm and higher probabilities of detection, simulation results show that the new GWO scheme performs better than the BBO method described in the literature in most cases.