A PERSYMMETRIC MULTIBAND GLR ALGORITHM

By exploring the covariance structure information to reduce the uncertainty in adaptive processing, a persymmetric generalized likelihood ratio algorithm (PGLR) is presented, together with the closed-form expressions of probabilities of detection and false alarm. The new multiband algorithm which requires less computation, can significantly outperform the corresponding unstructured multiband generalized likelihood ratio (GLR), especially in a severely nonstationary/nonhomogeneous interference environment. Furthermore, it is found via simulation that the constant false alarm rate (CFAR) performance of the new algorithm is as insensitive as the unstructured multiband GLR to the departure of interference distribution from Gaussian.