混沌神经网络编码研究

The Research on the Chaotic Neural Network Encoding

本文运用神经网络的非线性动力学理论，研究了利用线性分段的一波三折函数构造混沌神经网络，进行混沌序列编码。文中分别给出自环反馈网络模型和耦合网络模型，并针对每类模型，仿真得到了具体的网络，获得了相应的混沌编码。同时对耦合网络的模式识别进行了初步探讨。

In this paper, a new kind of transfer function of the neuron is defined as a piece wise linear N shaped (PWLN) function, which is a novelty that allows control of the collective chaotic neurons. The chaotic neuron model on PWLN is discussed and the chaotic behavior is verified with nonlinear dynamics of neural network. Then two kinds of chaotic network, namely feedback circles network and coupled network are proposed, which are used to chaotic coding. The latter is of great improvement to the former without reducing the coding capacity. Compared with other chaotic encrypting system, our networks have some advantages: (1) simple structure and clear topology; (2) high coding efficiency and great function; (3) it can not be decrypted by ways in existence.