FIR filter design: frequency-sampling method based on evolutionary programming

Abstract

Frequency sampling is one of the more usual methods in FIR digital filter design. In the frequency-sampling method, the values of transition-band samples, which are usually obtained by searching a table, must be determined in order to make the attenuation within the stop-band maximal. However, the value obtained by searching the table cannot be ensured to be optimal. Evolutionary programming (EP), a multi-agent stochastic optimization technique, can lead to globally optimal solutions for complex problems. In this paper, a new application of EP to the frequency-sampling method is introduced. Three examples of FIR filter design are presented, and the steps of EP realization and experimental results are given. The experimental results have shown that the values of transition-band samples obtained by EP can be ensured to be optimal and the performance of the filter is improved.