A Neuron-MOS Threshold Element with Switching Capacitors

In the paper a new approach to building a MOS threshold element and an arti cial neuron on its base is discussed. The implementability of the known MOS threshold elements is restricted by the sum of input weights. It is shown that by switching the element capacitors in the pre-charge phase it is possible to reduce the implementability restriction to the maximum value of the threshold. It essentially expands the set of threshold functions realizable on one element.

• Type: Conference
• Year: 2001
• Where: FUZZY
• Authors: Victor Varshavsky, Vyacheslav Marakhovsky
• Keywords: Arti Cial Neuron, FUZZY 2001, Fuzzy Logic, Mos Threshold Elements, Threshold Functions