混沌神经网络中的超混沌复杂性研究

Hyperchaos complication of chaotic neural network

在混沌保密通信中,混沌的复杂性是评估算法安全性的重要指标.通过对一种典型混沌神经网络的Lyapunov指数、相关函数、输出序列的频谱和功率谱等特征函数的综合分析,发现混沌神经网络中存在超混沌现象,并对其复杂性进行综合评价.同时,研究了参数变化对混沌神经网络Lyapunov指数谱的影响并进行了实验,结果表明,网络的输出状态可根据需要进行改变.

Complexity of chaos is an important index for evaluation of the security of algorithms in chaotic secure communication. Following a synthetical analysis on the key parameters of a typical chaotic neural network, such as Lyapunov exponent spectrum, correlative function, the frequency spectrum and power spectrum of output sequence, we found the existence of hyperchaotic phenomenon in the network. A comprehensive evaluation of its complication was made. The test was conducted to study the effect of the parameters on Lyapunov exponential spectrum of chaotic neural network. The results show that the output of network can be changed as demanded.