ASIC Implementation of Biologically Inspired Spiking Neural Network

Abstract

This paper presents the biologically inspired spiking neural network which can mimic the operation of our biological neural cell using CMOS implementation. Hardware implementation of a spiking neural network (SNN) shows more reliability and it comprises less area with high speed, which is much more than the neural network which has software counterpart. A body biasing technique has been used for the implementation of potassium and sodium ions in the neural cell. A fast neuron is required for the applications such as in image processing, data mining, supercomputing etc. Structural forms of spiking neural network and with synaptic functions are shown along with the simulation results. The main advantage of the proposed circuit is that it has a compact design, low power, and delay. In the proposed circuit power is decreased by 33.95 percent and delay is reduced by manifold. Power delay product (PDP) is approximately increased by 18× as a comparison with that of the reference circuit. © 2019 IEEE.