Digital filter design at gate-level using evolutionary algorithms

Traditionally digital filters are designed using the concept of a linear difference equation with the output response being a weighted sum of signal samples with usually floating point coefficients. Unfortunately such a model is necessarily expensive in terms of hardware as it requires many large bit additions and multiplications. In this paper it is shown how it is possible to evolve a tiny feed-forward rectangular array of logic gates to perform various filtering tasks lowpass, bandpass, and multiband. The circuit is evolved by assessing its response to digitised pure sine waves. Some of the evolved circuits possess almost linear properties, which means that they are capable of filtering composite signals which have not been encountered in training.