Analysis And Design Of Several Cryptographic Schemes Based On Complex Dynamic Systems

With the rapid development of computer science and communication technology, the society relies increasingly on information and information technology, which implies the advent of information era. However, information is prone to attacks such as eavesdropping, tampering, forgery and replay, the construction of information society is thus blocked. As a result, the security and protection of information have allured much attention from all walks of life.Cryptography constitutes the basis of information security. There is natural relationship between complex dynamic systems and cryptography: strong dynamical properties of complex system imply strong cryptographical properties, and the basic way to design strong cryptosystem implies complex dynamics. Until now, a large number of cryptosystems based on chaos and cellular automata have been proposed and systematically analyzed. While cellular neural network and complex network, which are promising for cryptography application, have also been theoretically studied.This dissertation focuses on the analysis and design of cryptosystems based on complex dynamic systems. The main achievements and originality contained in this dissertation are as follows:Some relevant concepts about cellular automata, cellular neural network and complex network are systematically presented. And the research status about chaotic and cellular automata-based cryptography is amply summarized.In the light of information theory, cryptography theory and the theory of complex dynamic systems, a novel chaotic encryption scheme based on arithmetic coding is proposed. The system is pollable, and has overcome the defect of chaotic behavior degradation in digitallization as well as the invariance of keystream when secret key is fixed. It also outperforms some previous chaotic cryptosystems in both performance and security. Furthermore, the arithmetic coding dose also contribute to the security of the proposed algorithm as a model of non-linear system.A fundamental defect of a class of image cryptosystem based on the replacement of the pixel values using recursive cellular automata substitution is found, and a novel chosen-plaintext attack scheme is presented. Some minor modifications and their influence on the security and performance of the cipher are also investigated. Based on the analysis of the keysteam generator in the original cryptosystem, we point out its defects and proposed a new pseudo-random number generator, which is provided with good random statistic and unpredictability.Some properties of cellular neural networks, which are fit for cryptography application, are discussed. Based on cellular neural networks designing, we found that it can be used as a platform to generate chaotic phenomena and realize arbitrary Boolean functions.According to the constraints of key distribution in wireless sensor networks, we present a novel key distribution scheme exploiting the synchronization of complex networks.Finally, the research work of this paper is summarized, and the future research direction is indicated.