Design of FIR digital filters using weighted peak-constrained least-squares optimization

Abstract

This paper is concerned with the application of the peak constrained least-squares (PCLS) optimization to the design of finite-impulse-response (FIR) digital filters. The salient feature of the PCLS optimization is that it permits the simultaneous min-max (MM) and least-squares (LS) optimization in a single run, leading to the design of digital filters exhibiting equal-ripple passband and minimum energy stopband magnitude-frequency response characteristics. The PCLS optimization proves to be particularly effective for the design of FIR digital filters having a large number of coefficients (both for symmetric as well as asymmetric digital filters), typically taking a small number of iterations to converge to the final optimal digital filter.