IIR filter synthesis based on Real-Coded Genetic Algorithm for human audiogram emulation

Abstract

This paper presents a method for optimized synthesis of IIR filters with user-defined magnitude response, based on a realcoded genetic algorithm. The application envisaged in this paper is the design of filters for hearing aids, which have to match given human audiograms. Two metrics for the differences between the synthesized and wanted magnitude characteristics are minimized by the optimization algorithm: the peak magnitude error and the total squared error. A design example is shown, with the IIR filter yielded by the proposed method for a particular audiogram compared against the FIR filter synthesized by using a previously reported method, for the same audiogram.I.

Acoustic Feedback Cancellation in Efficient Hearing Aids using Genetic Algorithm

“…The usage of real numbers allow us to extend the search domain by reducing the precision which has also the capacity to fine tune the solutions. Therefore comparing to binary coded strings, real numbers have more benefits in RCGA [7].…”

“…The RCGA is applied to solve the synthesis of digital filter whose magnitude response matches the human audiogram. The IIR filter exhibits significantly better performances than the FIR filter in approximating the required magnitude characteristic and in minimizing the resources necessary for the implementation [7]. The genetic algorithm is used with restricted search, is described by Enrique Alexandre [1], is explored for the above mentioned feature selection.…”

An artificial patient for pure-tone audiometry

“…To personalize hearing aids, Szopos et al synthesized in [27] human audiograms based on the Real-Coded Genetic Algorithm. Note that cross-hearing cannot be incorporated in their algorithm, and convergence of their algorithm is not guaranteed either.…”