Peak-constrained least-squares design of IIR digital filters

Abstract

An approach for approximating infinite duration impulse response (IIR) digital filters is proposed. The method is able to compromise maximum stopband attenuation and minimum stopband energy requirements. The approach is based on the weighted-least-squares (WLS) method with a weight function which is made constant within a given frequency interval at each iteration. In that sense, IIR digital filters with partially WLS-like and partially equiripple stopbands are efficiently designed.