Title :
Synthetic neural systems using current-mode circuits
Author :
Andreou, Andreas G.
Author_Institution :
Electr. & Comput. Eng., Johns Hopkins Univ., Baltimore, MD, USA
Abstract :
An approach to the engineering of VLSI analog computing systems is presented. The approach emphasizes the role of currents as an appropriate signal representation, the need for low power dissipation, and the need for simplicity in the basic functional circuits. The design methodology is based on current-mode (CM) MOS circuits, and is inspired by the functional and organizational principles of biological neural circuits. The basic circuit blocks exploit the translinear properties of the MOS transistor in the subthreshold and use unidirectional current signals. Circuits for associative memories and silicon retinas are described
Keywords :
MOS integrated circuits; VLSI; analogue computer circuits; content-addressable storage; neural nets; VLSI analog computing systems; associative memories; biological neural circuits; current-mode MOS circuits; current-mode circuits; functional circuits; power dissipation; signal representation; silicon retinas; subthreshold; translinear properties; unidirectional current signals; Analog computers; Associative memory; Current mode circuits; Design methodology; MOSFETs; Power dissipation; Power engineering and energy; Power engineering computing; Signal representations; Very large scale integration;
Conference_Titel :
Circuits and Systems, 1990., IEEE International Symposium on
Conference_Location :
New Orleans, LA
DOI :
10.1109/ISCAS.1990.112501