Application of an Improved Varying-Step Algorithm in the Real-Time Signal Filtering

This paper has applied wavelet transform domain adaptive filter (WTDFA) into real-time signal filtering, to accelerate the convergence velocity of filtering algorithm, and to avoid the disadvantages of wavelet transform (WT) and adaptive filter (AF). The paper also presents an improved varying step (IWS) algorithm based on WTDFA, which can be proved to ensure the convergence when the initial step factor satisfies a certain condition. The IWS algorithm can reconcile the contradiction between the velocity of initial convergence (VIC) and steady error (SE), and then the fastest VIC and smallest SE can be obtained in the filtering process. The simulation results verified the aforementioned conclusions.