Research on Blind Source Separation for Mainlobe Interference Suppression Based on Variable Step Size EASI Algorithm

This paper presents a novel method for suppressing mainlobe interference in radar systems using a variable step-size EASI algorithm. By analyzing the limitations of traditional BSS algorithms, we introduce a real-time step-size adjustment mechanism to improve the convergence speed and reduce steady-state errors. The proposed algorithm adaptively adjusts the step-size based on a separation degree metric, achieving optimal demixing and accurate source signal estimation. Simulation results demonstrate that the algorithm effectively suppresses noise amplitude modulation and SMSP interference, with a suppression improvement of over 40 dB for noise-AM interference and over 55 dB for SMSP interference. Compared with existing methods, the proposed algorithm achieves faster convergence and higher interference suppression accuracy.