一种改进的PWM型VLSI神经网络的设计

神经网络的超大规模集成电路 (VL SI)实现是发挥其优势的有效途径。改进了现有的基于脉宽调制 (PWM)技术的 VL SI神经网络设计方式。提出了一种结构简单的突触乘法器 ,它的精度高、线性范围大 ,而且不受开关噪声的影响。设计了一个增益可调的电压型 sigmoid变换电路 ,用以实现不同的神经元激活函数。提出一个 PWM所必需的电压-脉冲转换电路 ,它具有较高的转换精度和线性度。以这 3种电路为基础设计了一个解决异或 (XOR)问题的 PWM型VL SI神经网络。模拟结果表明其功能正确 ,具有较高的识别速度 ,适于神经网络的 VL SI实现

Design of a modified PWM VLSI neural network

The VLSI implementation of neural networks is an effective approach to facilitate its use. The existing VLSI neural network design method based on the pulse width modulation (PWM) technique is modified in this paper. A simple synapse multiplier is proposed, which has high precision, large linear range and less switching noise effects. A voltage mode sigmoid circuit with adjustable gain is designed for realization of different neuron activation functions. A voltage pulse conversion circuit required for PWM is suggested, which has high conversion precision and linearity. These 3 circuits are used to design a PWM VLSI neural network to solve the exclusive or (XOR) problem. The simulation result shows that it can function correctly with fast recognition speed; therefore, it is suitable for VLSI implementation of neural network.