On optimal convergence factor for IIR adaptive filters

In this work a variable convergence factor for use in Steiglitz-McBride (SM) based algorithms is proposed. The implementation of the convergence factor requires information available in the algorithm, therefore the additional cost is minimum. Some properties related to the performance of the variable convergence factors for the SM and output error methods are presented. Some simulations are also presented to confirm that the proposed convergence factor leads to fast convergence of the adaptive filter