基于Nios Ⅱ的脉冲神经网络硬件实现方法  

Method for hardware implementation of spiking neural network based on NiosⅡ

作　　者：张文娟[1] 王蕾[1] 王连明[1] ZHANG Wen-juan WANG Lei WANG Lian-ming(School of Physics,Northeast Normal University,Changchun 130024,China)

机构地区：[1]东北师范大学物理学院,吉林长春130024

出　　处：《东北师大学报（自然科学版）》2016年第4期57-62,共6页Journal of Northeast Normal University(Natural Science Edition)

基　　金：吉林省科技应用基础研究项目(20130102028JC);中央高校基本科研业务费专项资金资助项目(2412015KJ006)

摘　　要：将脉冲神经网络的理论和算法应用于函数拟合研究,通过使用CycloneⅡEP2C35F672C8N型FPGA芯片和基本外围电路,并基于NiosⅡ软核技术,建立了脉冲神经网络硬件模型.以指数函数为例拟合曲线,系统训练结束后稳定误差可达到0.2.实验结果表明,基于NiosⅡ的硬件实现方法能够成功地实现脉冲神经网络,为人工神经网络的研究提供了有效的仿真平台,同时该方法能够有效地模拟连续函数,扩展了神经网络的应用领域.Spiking neural network theory and algorithm are applied to fitting function in this paper, and spiking neural network model is founded based on Nios Ⅱ software using Cyclone Ⅱ EP2C35F672C8N and basic peripheral circuit. Exponential function is fitted and the stabilization error after training is 0. 2. The results show that hard implementation of spiking network based on Nios Ⅱ which can provide an effective simulation platform for the research of artificial neural network is successful. At the same time this method can simulate the continuous function effectively and expands the application fields of the neural network.

关 键 词：脉冲神经网络 硬件实现 NiosⅡ 曲线拟合 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]