A Research Of Radar Adaptive Detection Algorithm In Compound Gaussian Clutter

With the development of radar,the traditional Gaussian model at low-resolution radar is no longer suitable for time-varying and heavy-tailed sea clutter modeling at the high-resolution radar.A feasible method is to use a compound Gaussian model with texture components replacing the traditional Gaussian model.This dissertation studies radar tar-get detection problems by devising adaptive algorithms in compound Gaussian clutter to improve the target detection performance in the matched steering vector situation,the mis-matched rejection performance in the mismatched steering vector situation,and guarantee the constant false alarm rate(CFAR)property.Based on radar clutter model,mismatch model,polarimetric model,this dissertation uses persymmetric covariance matrix esti-mate,maximum a posteriori(MAP)estimate,generalized likelihood ratio test(GLRT),Rao test,Wald test,and CFAR detection to research detection algorithms.The main con-tributions and innovations of this dissertation can be summarized as follows:1.For the problem of mismatched steering vector target detection in compound Gaus-sian clutter with inverse Gaussian texture,the compound Gaussian mismatched detection algorithm is researched based on the adaptive beamformer orthogonal rejection test.Mod-eling the clutter as compound Gaussian distribution with inverse Gaussian texture,adap-tive mismatched detectors are proposed based on two-step GLRT and MAP GLRT criteria,respectively.The proposed mismatched detectors have the CFAR properties with respect to(w.r.t.)the real clutter covariance matrix,and have a good balance between the detection performance under the matched steering vector case and mismatched rejection properties under the mismatched steering vector case.2.For the problem of point-like target detection in compound Gaussian clutter with inverse Gaussian texture,adaptive CFAR detectors are devised based on two-step MAP Rao and MAP Wald criteria.In compound Gaussian clutter with inverse Gaussian texture,for the problem of sample-limited range-spread targets detection,adaptive persymmetric CFAR detectors are proposed based on two-step MAP GLRT,MAP Rao,and MAP Wald criteria,respectively.The proposed point-like target detectors improve in the detection performance and mismatched rejection performance,and range-spread targets detectors improve at least 2d B than non-persymmetric detectors in detection performance.3.For the problem of polarimetric target detection in compound Gaussian clutter with inverse Gaussian texture,three adaptive polarimetric inverse Gaussian detectors are devised based on the two-step method and MAP estimate.For the problem of polari-metric target detection in compound Gaussian clutter with Gamma texture,adaptive po-larimetric Gamma detectors are proposed based on combining the two-step GLRT,Rao,and Wald criteria and MAP estimate,respectively.For the problem of polarimetric target detection in compound Gaussian clutter with inverse Gamma texture,low computational-complexity adaptive polarimetric inverse gamma detectors are researched based on the two-step MAP GLRT,MAP Rao,and MAP Wald criteria,respectively.The proposed polarimetric detectors have CFAR properties w.r.t.texture statistics,and the proposed detectors improve averagely 1d B over their competitors in detection performance.4.In the compound Gaussian clutter environments,for the problem of sample-limited polarimetric target detection,adaptive polarimetric persymmetric(P~2)detectors are de-vised based on the two-step GLRT,Rao,and Wald criteria.In compound Gaussian clutter with inverse Gaussian texture,for the problem of sample-limited polarimetric target de-tection,adaptive P~2inverse Gaussian detectors are derived based on the two-step method and MAP estimate.In compound Gaussian clutter with Gamma texture clutter,for the problem of sample-limited polarimetric target detection,adaptive P~2Gamma detectors are proposed based on combining the two-step GLRT,Rao,and Wald criteria and MAP esti-mate.In compound Gaussian clutter with inverse Gamma texture clutter,for the problem of sample-limited polarimetric target detection,low computational-complexity adaptive P~2inverse Gamma detectors are researched based on the two-step MAP GLRT,Rao,and Wald criteria.The proposed P~2detectors have CFAR properties w.r.t.texture statistics and the real polarimetric clutter covariance matrix,and improve averagely 1d B over their competitors in detection performance.5.For the problem of polarimetric mismatched steering matrix target detection in compound Gaussian clutter,adaptive polarimetric mismatched CFAR detectors are de-vised based on the quasi-whitened space and truly whitened space,which improves in detection and mismatched rejection performance.For the problem of target detection in different compound Gaussian scenarios,the matched detection schemes are devised,mak-ing the above-proposed detectors integrated into the united framework system of detection.