多阈值神经元电路设计及在多值逻辑中的应用

Circuits Design of Multi-Thresholded Neuron(MTN) and Its Applications in Mult-Valued Logic

分析了多阈值神经元工作原理,并提出设计多阈值神经元电路的方法.首先,用两个 MOS晶体管组成电压型突触电路,然后又提出一种基于BiCMOS工艺的判别转换开关电路,这种电路以压控电流作为阈值信号,并实现电压到电流的转换.在此基础上,结合限幅电压开关理论提出多阈值神经元阈值判别函数电路的开关级设计方法.最后,从开关级设计了实现三值逻辑中文字、与、或三种基本运算的多阈值神经元电路,用这三种基本运算的多阈值神经元电路可实现任意三值函数的多阈值神经网络.文章还对设计出的电路用 PSPICE进行模拟,测量相关参数.模拟结果表明,该文设计的电路不仅实现了正确的逻辑功能,而且速度较快.

By analysing the principle of muti-thresholded neurons (MTNs), a method was developed for designing muti-thresholded neuron circuits (MTNCs). First, the voltage-mode synapse circuit was designed using two MOS transistors. Second, a BiCMOS technics based circuit called verdict-converting switch (VCS) was put forward. The VCS has a threshold which is voltage-controlled current signal and it has the ability of converting the voltage signal to current signal. Based on this possibility, an approach was proposed for designing the circuits of multi-thresholded verdicting function (MTVF) at switch level. Finally, several MTNCs were designed for implementing the Literal, AND, OR operation as three basic operations in ternary logic at switch level. With these basic MTNCs, arbitrary ternary functions can be achieved by multi-thresholded neural networks. The result of simulation with PSPICE showed that the designed circuits had not only the correct logic function but also had small propagation delay.