基于混沌控制系统的神经网络异步加密  

Asynchronous Encryption of Neural Networks Based on Chaotic Control System

作　　者：邹阿金[1] 肖秀春[1] 

机构地区：[1]广东海洋大学信息学院,湛江524088

出　　处：《计算机工程》2008年第12期160-161,共2页Computer Engineering

摘　　要：将Chebyshev神经网络模型作为混沌控制系统辨识器,任选系统初值和非线性、非周期性控制律,通过Chebyshev混沌神经网络产生混沌序列,求出该序列的排序置换及逆置换,对明文置换后即可得密文,密文逆置换即可还原得到明文。理论分析和加密实例表明,该加密算法可异步进行,且只需改变系统初值和控制律,便可获得不同的混沌序列,从而实现"一次一密"加密,具有良好的安全性。This paper introduces new feed-forward neural networks where hidden units employ orthogonal Chebyshev polynomials for their activation functions and uses it as the identifier of the chaotic control system.Through the Chebyshev neural networks,it can produce chaotic series by varying system initial value and nonlinear,nonperiodic control rule,calculate the sort permutation and inverse permutation of this series,and permute the plaintext to achieve ciphertext,inverse permute the ciphertext to get the plaintext.Theoretical analysis and encryption instances prove that the arithmetic can implement asynchronous one-time pad cipher encryption by varying the system initial value and control rate.

关 键 词：神经网络 混沌 置换 异步加密 控制系统 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]