On dynamic behavior of full range CNNs

Cellular neural networks are described by large sets of coupled nonlinear differential equations. Two CNN models were proposed in the literature: the original Chua-Yang model and the Full-Range model, that were exploited for VLSI implementations. We show that, even if the two models exhibit similar properties, they are not topologically equivalent, i.e. there exist sets of identic parameters for which they present a qualitatively different dynamic behavior. As a consequence, all the results based on the original Chua-Yang model, should be carefully checked, before extending them to the Full-Range model.