混沌序列的神经网络实现  

Chaos generation based on neural network

作　　者：石文孝[1] 荆涛[2] 杨怀江[2] 

机构地区：[1]长春邮电学院通信工程系,吉林长春130012 [2]中国科学院长春光学精密机械与物理研究所,吉林长春130021

出　　处：《光学精密工程》2000年第3期231-233,共3页Optics and Precision Engineering

基　　金：国家自然科学基金资助项目; 邮电部重点科技发展计划资助项目

摘　　要：利用ＢＰ神经网络，对非线性系统产生的混沌序列进行学习，逼近该非线性系统的映射特征，从而使自身具有混沌特性，成为具备混沌输出能力的神经网络，建立了基于神经网络的混沌序列产生模型（ＣＧＮＮ）。此方法利用具有学习能力的神经网络，它不需要针对某一种非线性映射设计单一的系统结构．并利用ＤＳＰ技术在ＣＧＮＮ基础上，制成了混沌神经网络协处理机插板，将其产生的混沌序列作为密钥，用于信息的保密通信。Back propagation neural network and chaos dynamics are combined in this paper. The global-optimum and dynamic-adjusting neural network is trained by the chaos series from non-linear system so as to approach the mapping feature of that system. So, a chaos generation scheme based on the neural network is proposed. By using nonlinear neural network with learning ability, this scheme doesn't have to be designed to be a single system structure for a certain nonlinear chaotic map. At last, using DSP technology, a Chaos Neural Network Co-processer Flashboard is made based on the CGNN. The chaos is used as a encrypeted key in the encryption telecommunication.

关 键 词：混沌 BP网络 DSP CGNN模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN911.6[自动化与计算机技术—计算机科学与技术]