IIR adaptive filters using discrete orthogonal functions

A novel adaptive IIR (infinite impulse response) system identification algorithm is described. It is based on resolving the target function into a set of discrete orthogonal functions which in turn ensures complete independence of subsequent sections. Examples show that this method always results in a much improved adaptation rate compared with the other approaches. It is concluded that using discrete orthogonal functions improves significantly the convergence properties of adaptive recursive filters.<
>