Nyquist Filter Design using POCS Methods: Including Constraints in Design

Abstract

The problem of constrained finite impulse response (FIR) filter design is central to signal processing and arises in a variety of disciplines. This paper surveys the design of such filters using Projection onto convex sets (POCS) and discusses certain commonly encountered time and frequency domain constraints. We study in particular the design of Nyquist filters and propose a simple extension to the work carried out by Haddad, Stark, and Galatsanos in [1]. The flexibility and the ease that this design method provides in terms of accommodating constraints is one of its outstanding features.