摘要
提出了一种基于SoPC的神经网络的硬件实现方法,该方法以FPGA器件为硬件载体,NIOSII软核处理器为CPU,Avalon片内总线为数据交换架构。研究了多层前馈神经网络在FPGA上的实现方法,描述了神经网络模块与Avalon片内总线的接口技术。整个系统在Altera的EP2C8Q208C8器件上实现,结果表明,该方法的应用不仅提高了人工神经网络的运算速度,还提高了整个系统的灵活性。
This paper reports on the implementation of an artificial neural network based on the SoPC. This method integrates a soft processor which named NOIS II and an in-chip buss named Avalon to a single FPGA device. It studies the implementation of the multilayer artificial neural network on FPGA and describes the interface of neural network and Avalon bus. This system is implemented with EP2C8Q208C8 of Altera. It is show that the method not only step up the operation but also increase the flexible of artificial neural network.
出处
《电子测量技术》
2009年第6期116-118,123,共4页
Electronic Measurement Technology
