Design of stable minimax IIR digital filters using semidefinite programming

Semidefinite programming (SDP) has recently been found useful in designing various types of FIR digital filters. This paper describes a SDP-based method for the design of stable minimax IIR filters. A single linear matrix inequality (LMI) constraint assures the filter's stability and fits nicely into the SDP-based design setting. Design efficiency and performance of the proposed method an illustrated by simulations and comparisons to a well-known design by Deczky (1972).