基于混沌神经网络的字序列密码研究与设计

发布时间：2018-05-26 00:21

本文选题：混沌神经网络 + 混沌吸引子　；参考：《华侨大学》2013年硕士论文

【摘要】：随着信息通信技术的发展与普及，信息安全的问题也层出不穷，如何保障国家安全和国民经济正常运行显得尤为重要。序列密码是一种非常重要的密码体制，在信息安全领域发挥着重要的作用。如何设计具有自主知识产权的新型高性能序列密码算法，一直是信息安全研究的热点。尽管传统的经典算法在信息安全的保密方面发挥了重要的作用，但随着密码分析技术的发展和具有强大计算能力的计算机出现，，使得一些传统的经典算法受到了极大的冲击，迫使设计者开始把目光投向新型的加密算法。混沌神经网络具有良好的随机性、遍历性、混合性、确定性和对初值参数的敏感性以及高度并行性运算，特别适合用于加密。本文首先学习了密码学和混沌学的基础理论知识、不同的加密体制和几种混沌神经网络模型，重点学习了Hopfield模型的非线性动力学特性。通过对eSTREAM Project中的各种算法和传统加密系统中的各种部件的总结，提出了两种基于OHNN的序列密码加密算法。其一：基于OHNN和M-LFSR的序列密码加密算法，该算法利用工作在过饱和状态下的离散的Hopfield神经网络所生产的混沌吸引子对改进型线性反馈移位寄存器的进行非线性选择输出，完成流加密。其二：基于OHNN和驱动表的序列密码加密算法。该算法同样是利用工作在过饱和状态下的离散Hopfield神经网络所产生的混沌吸引子对驱动表中的经过变化的随机数进行非线性选择输出，完成流加密。最后，在VC++6.0的平台上，实现了两种算法的软件仿真，在Matlab和RedHat9.0平台上，完成了对两种算法的各种测试。通过测试和安全性分析表明，这两种算法具有良好的安全性和加解密速度，为保密通信提供了新选择。
[Abstract]:With the development and popularization of information and communication technology, the problems of information security emerge in endlessly. How to ensure national security and the normal operation of national economy is particularly important. Sequential cryptography is a very important cryptosystem, which plays an important role in the field of information security. How to design a new high performance sequence cipher algorithm with independent intellectual property has been a hot topic in information security research. Although the traditional classical algorithms play an important role in the security of information, with the development of cryptographic analysis technology and the emergence of powerful computer, some traditional classical algorithms have been greatly impacted. It forces designers to start looking at new encryption algorithms. Chaotic neural networks have good randomness, ergodicity, mixing, certainty, sensitivity to initial parameters and high parallelism, so they are especially suitable for encryption. In this paper, we first study the basic theory of cryptography and chaos, different encryption systems and several chaotic neural network models, focusing on the nonlinear dynamics of the Hopfield model. By summing up various algorithms in eSTREAM Project and components in traditional encryption system, two kinds of sequential encryption algorithms based on OHNN are proposed. First, a sequence cipher encryption algorithm based on OHNN and M-LFSR, which uses chaotic attractor produced by discrete Hopfield neural network working in supersaturated state, performs nonlinear selective output of the improved linear feedback shift register. Complete stream encryption. Second, the sequence cipher encryption algorithm based on OHNN and driver table. The chaotic attractor generated by the discrete Hopfield neural network working in supersaturated state is also used to perform the nonlinear selection output of the variable random numbers in the driving table to complete stream encryption. Finally, the software simulation of two algorithms is realized on the platform of VC 6.0. On the platform of Matlab and RedHat9.0, various tests of the two algorithms are completed. The test and security analysis show that the two algorithms have good security and encryption and decryption speed, which provides a new choice for secure communication.
【学位授予单位】：华侨大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TN918.4

【相似文献】