CMOS based β-driven threshold elements with functional inputs

We describe a CMOS implementation of a β-driven threshold element. Using such an element as a functional basis of artificial neurons has a number of advantages in comparison with the traditional functional basis. In this paper we focus on extension of the functional capability of the threshold element. We introduce functional inputs into the β-driven threshold element to allow increase of its functional capability up to the capability of realizing every monotonic Boolean function. We present an implementation of the β-driven threshold element, having the proposed functional inputs, by a CMOS circuit. SPICE simulation results demonstrate the high efficiency of the solutions and consequently motivate further investigations of artificial neurons based on β-driven elements.