一种神经元模拟电路的分析和应用

Analysis and Application of An Artificial Neuron Analog Circuit

本文研究神经元的模拟电路实现，分析一种可编程突触电路、变增益细胞体电路，用它们构成一个８×７单层感知机网络。以３μｍＣＭＯＳ工艺第二类模型参数进行ＰＳＰＩＣＥ模拟，实现了预期的模式识别功能。

An analog circuit hardware implementation of artificial neuron is studied. The programmable synapse circuit and gain adjustable neuron circuit are analyzed. As an application example,a 8 X 7 single-layer perceptron network is built. PSPICE simulation results using level-2 model parameters of a 3μm CMOS process are given,and the predicted function for pattern identification is obtained.