Rounding FIR Filter Coefficients Using Genetic Algorithms
The quantisation process of filter coefficients required due to the finite length of the register provided to store the coefficients is an important step in the realization of a given filter. This process must be optimally carried out to minimise the error in the filter response caused by the quantisation of coefficients. Genetic algorithm is a heuristic optimisation procedure which is able to optimise both discrete and continuos functions efficiently. This paper explains how to use a genetic algorithm to round the coefficients of an FIR (finite-impulse-response) filter represented using fixed-point numbers and presents the simulation results.
KeywordsGenetic Algorithm Filter Coefficient Stopband Edge Desire Frequency Response Hyper Signal
Unable to display preview. Download preview PDF.
- 1.Jenkins, W.K. and Leon, B.J. An analysis of quantisation error in digital filters based on interval algebra’s, IEEE Trans. Circuits Syst., vol. Cas-22, 223, (1975).Google Scholar
- 2.Holland, J.H. Adaptation in natural and artificial systems, Ann Arbor: University of Michigan Press, USA, 1975.Google Scholar
- 3.Proakis, J.G. Digital signal processing (Principles, algorithms, and applications) (Sec. ed.), Macmillan Publishing Company, New York, 1992.Google Scholar
- 4.Antoniou, A. Digital Filters (Analysis, design, and applications), (Sec ed), McGraw-Hill, Inc., USA, (1993).Google Scholar
- 5.Davis, L.(ed.) Handbook of genetic algorithms, Van Nostrand Reinhold, NY, New York, 1991.Google Scholar
- 6.Karaboğa N., Wani A. and Karaboğa D. Optimisation of multivariable functions, Proc. of the Third. Conf. on Information Technology and its Applications, ITA’93, Markfield Conf. Centre, Leichester, April, 14–16, UK, 1994.Google Scholar
- 7.Hyper signal DSP software, (Users manual, Reference guide, Hardware and interface source code), Hyperception, Inc. 9550 Skillman LB 125, Dallas, 1991.Google Scholar