A genetic algorithm for optimisation of linear phase FIR filter coefficients


Oner M.

32nd Asilomar Conference on Signals, Systems and Computers, California, United States Of America, 1 - 04 November 1998, pp.1397-1400 identifier

  • Publication Type: Conference Paper / Full Text
  • City: California
  • Country: United States Of America
  • Page Numbers: pp.1397-1400
  • Middle East Technical University Affiliated: No

Abstract

A genetic algorithm is used to design and optimise digital FIR filter coefficients. Given the desired amplitude response of the filter to be designed, algorithm generates the filter coefficients with the specified number of taps and bits per coefficients. The linearity of the phase response is satisfied by making filter coefficients symmetric. Algorithm generates a population of genomes that represents the filter coefficients and compares the amplitude response of each genome to that of the desired amplitude response. New genomes are generated by crossover, mutation operations as well as deterministic pruning method. Since the algorithm directly generates digital coefficients, there is no need to truncate coefficients for digital hardware implementation of the filter.