Statistical analysis of an unbiased plain gradient algorithm for a constrained adaptive HR notch filter

Theoretical analysis of the unbiased plain gradient (UPG) algorithm for a constrained adaptive HR notch Alter (ANF) is presented in this paper. The difference equations for the convergence to the mean of the estimated parameter, estimation bias, and stability bound are derived. It is shown that the UPG provides bias smaller that that of the PG. Computer simulations have been treated to corroborate the theoretical analysis. © 2008 IEICE.