The Method Of Signal Detection Based On Revised Conjugate Gradient Stap

It is necessaryto obtain signals in space-time correlated environment, since directional jammers and space-time correlated clutter is unbalanced random process. This paper presents for the first time the theory of CG-AMF, and then proposes RCG to obtain the optical weight of STAP in Krylov subspace inlower dimension, and three RCG-AMF are discussed which employ a better initial vector of AMF, and reduce the condition of the covariance matrix and so on. Meanwhile, the excellent performance of CG-AMF, lower computational cost and better detection performance are got.On the one hand, RCG-AMF is based on the extreme eigenvalue. Lanczos method is applied to solve the Wiener-Hopf equations, an approximate weight of STAP and the largest(or smallest) eigenvalue of the covariance matrix can be obtained. So the initial vector of CGis optimized, and then CG is used to deal with the equations which incurs a series of RCG-AMF. Furthermore, the results show that the condition number is reduced to no more than 0.25 times(or about 0.25 times) and the spectral distribution is enhanced more than 6dB(or about 6dB) through a few of Lanczos steps.On another hand, RCG-AMF is based on Marcenko-Pastur spectral. The Marcenko-Pastur spectral is applied to estimate the spectral distribution of the covariance matrix, which is employed to get the extreme eigenvalues. Then, the reduced condition number of the Wiener-Hopf equations and a series of RCG-AMFcan be obtained. It shows that the condition number of the matrix is reduced to no more than 0.25 times the original and the spectral distribution is enhanced about 6dB.The methods offer better optical vectors and lower condition number, which lead a better detector and lower computational cost. The simulations and measured data have verified the efficiency, which serve as the basis for the real-time processing of STAP.