An adaptive detection algorithm with persymmetric covariance structure

By exploring the covariance structure information to reduce the uncertainty in adaptive processing, a persymmetric generalized likelihood ratio (PGLR) algorithm is presented, together with the closed-form expressions of its probabilities of detection and false alarm. The algorithm, which has a faster convergence rate and requires less computation, can significantly outperform the corresponding unstructured GLR, especially in a severely nonstationary/nonhomogeneous interference environment. It also possesses a constant false-alarm rate feature of practical importance