节点文献
基于DSP的BP神经网络的硬件实现
Hardware Implementation of DSP Based BP Neural Network
【摘要】 为了提高人工神经网络的运行速度,本文提出了采用多片TMS320C40构成的并行系统实现BP神经网络的具体方案,详细探讨了影响系统性能的各个因素,给出了系统性能的理论值,并且经过软件模拟证实了方案的可行性。
【Abstract】 This article Provoides a new method based on several TMS320C40s to implement BP neural net work. raising the trainin speed. All factor affecting the system performnce are discussed in detail. at the same time the fomular about this system is given and the practicibility of system is tested by software simiulation.
【关键词】 人工神经网络;
并行算法;
硬件实现;
【Key words】 Artificial Neural Networks; Parallel Algorithm; Hardware Implementation;
【Key words】 Artificial Neural Networks; Parallel Algorithm; Hardware Implementation;
- 【文献出处】 信号处理 ,SIGNAL PROCESSING , 编辑部邮箱 ,1998年02期
- 【分类号】TP18
- 【被引频次】15
- 【下载频次】261