A dynamic threshold value control method for chaotic neural networks

Abstract

A dynamic threshold value control method is proposed to get control over a chaotic neural network (CNN). The sinusoidal signal, which constitutes the key component of the brain wave, is taken as the control signal to change the threshold value of the internal states of the CNN. The chaos control of the CNN is therefore reached with its outputs of the controlled CNN reciprocating between a stored pattern and its reverse pattern related with the initial pattern. And then the CNN can be applied in information processing, such as pattern recognitionetc. The chaos control method works in a self-adaption way since it does not need to set the threshold value beforehand, which is in accordance with brains’ thinking activities.