A Tensor-Based Underdetermined Blind Identification Algorithm of Compound Radar Jamming Signals
Jiaqi Ruan, Mingzhang Zhou, Ruiping Song, Kun Ye, Haixin Sun · 2024
In order to convert compound radar jamming signals into individual jamming signals, while improving the accuracy of blind identification, this paper presents a novel underdetermined blind identification approach resorting to tensor decomposition. First, the autocovariance matrices for each individual sub-block are determined and subsequently stacked to form the third-order symmetric tensor of observed signals. Then, an optimal step proximal alternating least squares (OS-PALS) method with a self-adaptive step size selection rule is proposed to decompose the constructed tensor, which guarantees the convergence and accelerates the convergence rate in tensor decomposition. Finally, singular value decomposition (SVD) is used to eliminate the influence of error factors. Simulation results demonstrate that the proposed approach surpasses existing methods in regard to blind identification performance in common application scenarios.