三维Hopfield型神经网络的混沌同步研究

Study on chaotic synchronization for 3D Hopfield neural network

利用状态观测器理论提出了具有多个非线性项的混沌系统的标量输出同步方法。通过对误差系统的变换,以降低非线性项对误差系统的影响,从而可减少极点配置参数。基于这种方法,设计出含有多个非线性项的Hopfield型混沌神经网络的同步系统,并计算出同步参数,从而实现了含有多个非线性项的混沌系统同步,并用计算机仿真证实了此方法的有效性。Hopfield型混沌神经网络具有全局吸引的混沌吸引子,混沌性能稳定,采用线性电路器件容易实现,对混沌的实际应用有参考价值。

This paper proposes an observer-based method to design synchronization systems for the chaotic systems containing multiple nonlinear parts by a scalar output. By a transform on the error system, the dynamical effect from the nonlinear parts is reduced, and therefore the number of parameters for pole assignment is decreased. This method was successfully applied to the synchronization of the 3D Hopfield neural network which has three nonlinear parts. Computer simulation shows that the synchronization error becomes zero very fast, which suggests that the method is effective. Since the Hopfieldtype chaotic neural network has a global attractive attractor, this work may be used in practical application of chaos.