论文部分内容阅读
作为继电阻、电容、电感三大电路基本元件外的第四类电路元件忆阻器,可模拟大脑突触神经网络的存储记忆功能。构建一个传统三维细胞神经网络,利用磁控忆阻器的非线性特性替换传统细胞神经网络的输出模块。采用Multisim通用电路元件构建磁控忆阻等效电路,在电路整体设计上简化了输出函数模块数量,与具有混沌行为的CNN系统相比,新型忆阻CNN电路不仅展现出了混沌吸引子现象,而且忆阻内部的磁能量随细胞状态而变化,可完全达到等效输出函数的忆导值。数值计算与电路仿真结果验证了忆阻细胞神经网络的混沌特性及新设计方法的可行性,在信号处理、同步控制与图像加密等方面具有现实的应用价值。
As the resistance, capacitance, inductance of the three basic circuit components outside the memorizer of the fourth circuit components, can simulate the memory function of the brain synaptic neural network. Construct a traditional three-dimensional cellular neural network and replace the output module of the traditional cellular neural network with the nonlinear characteristic of the magneto-resistive memristor. Compared with the CNN system with chaotic behavior, the new memristive CNN circuit not only shows the phenomenon of chaotic attractor, but also can reduce the number of output function modules in the overall design of the circuit by using Multisim universal circuit components to build the equivalent circuit of magneto- Moreover, the internal magnetic energy of the memristor changes with the state of the cell and can completely reach the recuperative value of the equivalent output function. The numerical calculation and circuit simulation results verify the chaotic characteristics of the memristive cellular neural network and the feasibility of the new design method. It has practical value in signal processing, synchronization control and image encryption.