Two improved wavelet transform domain LMS sign adaptive filter algorithms against impulsive interferences (Q2118691)

Different wavelet transform (WT) domain least mean squares (LMS) WTDLMS algorithms have been studied. The current research extends the affine projection approach (APA) and sign adaptive algorithms to WTDLMS so as to propose the improved WTDLMS algorithm and its variants, which have better performances than the conventional WTDLMS. Concretely, the improved wavelet transform domain least mean squares (IWTDLMS) adaptive algorithm is established by extending the affine projection approach (APA) to WTDLMS algorithm, which could speed up the convergence rate of WTDLMS, especially for colored input signals. The IWTDLMS sign algorithm (IWTDLMS-SA) is proposed based on l$_1$-norm minimization to improve the convergence speed of IWTDLMS against impulsive interference. The variable step size IWTDLMS-SA (VSS-IWTDLMS-SA) is further established to improve the performance of IWTDLMS-SA in achieving a fast convergence speed and low steady-state error. The MVSS-IWTDLMS-SA is introduced to increase the tracking ability of VSS-IWTDLMS-SA. For the proposed algorithms, their computational complexity and convergence analysis are described, and their usefulness against impulsive noises is demonstrated through different simulation results.