Study Of Hyperchaos Delayed Neural Networks' Complexity To Secure Communication

The encryption is the key problem of secure communication. That applying chaos theory to secure communication and information encryption has already become one of the hottest research projects on the combination of nonlinear science and information science, and it is a novel branch of high-tech research fields. Neural networks, with the properties of nonlinear dynamics such as chaotic behavior and parallel processing, are regarded as one of good options for encryption algorithm applied in the secure communication.This paper introduced the delayed neural network hyperchaos theory to the modern information security and secure communications. Based on the complexity of hyperchaos properties evaluation of the delayed neural network, the paper conducted a comprehensive study for building a new generation of secure communication mechanism.In order to analyze the hyperchaos complexity of the delayed neural network synthetically and comprehensively, the author used evaluation method that integrate characteristic function analysis and randomness testing. First, in order to provide the basis to the chaotic complexity aims at the neural network in the analog communication application, the author analyzed the Lyapunov exponent of several common chaotic neural network models such as Chen&Aihara chaotic neural network model. Wang&Simth chaotic neural network model and delayed chaotic neural network models such as Mackey-Glass, Liao. Ikeda neural networks.Then aims at the neural network in the digital communication application, a set of randomness tests described by NIST (National Institute of standards and Technology) are used for the statistical analysis of binary chaotic and delayed chaotic neural network sequence.At last, for viewing the chaos sequence's frequency characteristic more directly and supplying the complexity analysis of the delayed neural network, the author used Periodogram and Burg spectrum analytic method to numerical value simulate and contrast the characteristic's power spectrum.