本文已被：浏览 2386次 下载 5289次
Received:July 17, 2013 Revised:August 19, 2013
Received:July 17, 2013 Revised:August 19, 2013
中文摘要: 脉冲神经网络被誉为第三代神经网络，近年来受到许多学者的关注，其优势已经在模式识别、计算机视觉等诸多领域得到了发挥。脉冲神经网络的硬件化是实现其强大计算能力的重要途径，而突触的硬件实现又是其中的一个关键性环节。本文先从SRM模型中脉冲神经元突触的特性曲线入手，用适合FPGA实现的差分方程来逼近这一响应曲线，并对差分方程中的待定参数进行了优化，然后根据差分方程，在Simulink平台上设计出硬件电路，并给出了在方波脉冲激励下，电路输出的仿真结果，最后对今后的工作做出了展望。
Abstract:Spiking neural network is considered as the third generation of neural networks, and it has attracted many researchers. Its advantages have been shown in pattern recognition and computer vision. The implementation of the spiking neural network in the hardware is an important method to show its powerful computation ability. This paper begins with the synapse response curve, and then in order to suit for the FPGA implementation the difference equation is used to approach the response curve. In addition, the genetic algorithm is used to optimize the parameters of the circuit. According to the difference equation, the circuit is designed in the Simulink platform. The simulation results are obtained for the outputs of circuit triggered by a square impulse wave. Finally the future work is discussed.
keywords: spiking neutral network synapse response curve Simulink FPGA
文章编号： 中图分类号： 文献标志码：
基金项目:福建省产学科技重大项目(2013H61010023)
引用文本：
李宏伟,吴庆祥.脉冲神经网络中神经元突触的硬件实现方案.计算机系统应用,2014,23(2):16-21
LI Hong-Wei,WU Qing-Xiang.Implementation of the Synapse of Spiking Neural Network in the Hardware.COMPUTER SYSTEMS APPLICATIONS,2014,23(2):16-21
李宏伟,吴庆祥.脉冲神经网络中神经元突触的硬件实现方案.计算机系统应用,2014,23(2):16-21
LI Hong-Wei,WU Qing-Xiang.Implementation of the Synapse of Spiking Neural Network in the Hardware.COMPUTER SYSTEMS APPLICATIONS,2014,23(2):16-21
