基于正弦映射的动态递归网络产生混沌序列实现方案  

A scheme of generating chaotic sequences by a dynamical recurrent neural network based on sine-mapping

作　　者：栾宇[1] 张兴周[1] 任艳春 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《应用科技》2006年第6期24-27,共4页Applied Science and Technology

基　　金：哈尔滨工程大学基金资助项目(No.HEUF04106);基础研究项目培育基金资助项目(No.HEUFP05024)

摘　　要：提出了应用正弦映射来产生混沌序列,正弦映射是结构简单的一维映射,通过对此映射的分析,得出其能在较大参数范围内产生混沌,并且产生的混沌序列具有较大的Lyapunov指数,具有比较丰富的动态特性,并将其引入到动态递归网络的隐含层作为非线性函数,选择合适的网络权值,便可在输出层产生具有复杂动力学行为的混沌序列.通过对此序列进行Lyapunov指数和相关性的仿真分析,得出其具有良好的随机性,可应用于混沌保密通信等领域.A method of generating chaotic sequences is proposed by using sine-mapping which is simply a one-dimensional mapping. This kind of mapping can be chaotic in a wide parameter range and the sequences generated possess larger Lyapunov exponent and excellent dynamical properties. The mapping is also introduced into the hidden layers of dynamical recurrent neural network as a non-linear function. With appropriate network weights selected, the output layers of the network can generate chaotic sequences with complex dynamical behaviors. By analyzing the simulation results to the Lyapunov exponent and the correlation, it is known that the sequences possess fine randomness, and can be applied in chaotic secure communication, etc.

关 键 词：混沌序列 动态递归神经网络 LYAPUNOV指数 一维映射 

分 类 号：TN913[电子电信—通信与信息系统]