Adaptive Detection of Point Targets in Compound-Gaussian Clutter With Inverse Gamma Texture
Qing Wang, Xuan Zhou, Weijian Liu, Jun Liu, Xueli Fang, Daikun Zheng
Abstract
This paper investigates the problem of adaptive detection of point targets against compound-Gaussian (CG) clutter with inverse gamma (IG) texture. Under the subspace signal model, we first propose adaptive detectors without signal mismatch based on the one-step generalized likelihood ratio test (1S-GLRT), maximum a posteriori Rao (MAP-Rao), MAP-Wald, MAP-Gradient, and MAP-Durbin criteria. These detectors have a unified expression, denoted as IG subspace-based GLRT (IG-SGLRT), which performs better than existing detectors. Then, in the case of mismatched signals, we add virtual interference under the null hypothesis and introduce a unified detector based on 1S-GLRT, MAP-GLRT, and two-step GLRT (2S-GLRT), named IG subspace-based whitened adaptive beamformer orthogonal rejection test (IG-SWABORT), which has better selectivity than the IG-SGLRT. Finally, on the basis of the above two types of unified detectors, we design a dual-parameter tunable detector, named tunable IG SGLRT WABORT (T-IG-SGWABORT), which can provide superior robustness or selectivity for mismatched signals by adjusting parameters. In addition, the tunable detector can even obtain better detection performance in the absence of signal mismatch than the existing detectors. Experimental results of simulations and measured data verify the effectiveness of the proposed detectors. Simulation experiments also demonstrate that all the proposed detectors have constant false alarm rate (CFAR) properties with respect to the covariance matrix.