CFAR Detector Performance in Nonhomogeneous Interference Environments

Abstract

Wide-area search radar systems must reliably detect targets over a large region of interest while maintaining a low, constant false alarm rate. Constant false alarm rate (CFAR) detectors have been proposed in the literature, and they operate by estimating statistical parameters of interference in the received signal to calculate a threshold that maintains a constant probability of false alarm. However, the nonhomogeneous nature of surface clutter degrades CFAR detection performance, often increasing the false alarm rate and reducing the probability of target detection. In this thesis, a simulation was developed to generate data representative of that produced by a ground-based surveillance radar. The simulation includes a typical processing chain employed by radar systems, and the detection performance of several CFAR detectors, including Cell-Averaging (CA), Greatest- of-Cell-Averaging (GOCA), Smallest-of-Cell-Averaging (SOCA), and Ordered-Statistics (OS), was analyzed using the simulated radar data. The OS detector was proven to be the optimal detector of the study, as it maintained the lowest false alarm rate and the highest probability of target detection when the simulated radar data contained nonhomogeneous interference.