Main-Lobe Interference Suppression Using Robust Independent Component Analysis Based on Complex Newton Optimization

Mainlobe interference suppression is a challenging problem in radar signal processing, requiring not only effective interference suppression but also maintaining inter-pulse phase continuity. However, current interference suppression methods, such as independent component analysis (ICA)-based methods, generally introduce phase distortions. In this paper, to ensure the phase continuity of target signal over the coherent integration time, we propose a mainlobe interference suppression method, that is robust independent component analysis based on complex Newton optimization (CN-RobustICA). The proposed CN-RobustICA introduces the complex Newton optimization method to estimate the separating vector of interference suppression and derives the globally optimal step size to prevent the separation vectors from falling into local optima. Therefore, CN-RobustICA can maintain both amplitude and phase consistency of the separating vector while preserving separation accuracy. Experimental results on real mainlobe interference datasets, including deceptive and suppressive jamming, demonstrate that CN-RobustICA successfully achieves mainlobe interference suppression and significantly improves inter-pulse phase stability compared to the recently proposed ICA-based methods. Meanwhile, because of the definition of a simplified kurtosis formulation, the proposed CN-RobustICA reduces computational complexity compared to the traditional RobustICA.