基于忆阻器的混沌系统设计及应用研究

发布时间：2018-03-30 14:08

本文选题：忆阻器　切入点：忆阻混沌系统　出处：《西南大学》2017年硕士论文

【摘要】：在信息化社会迅猛发展的今天,提高信息传递的保密性、有效性及效率成为当下科学和技术发展的重心。纳米级尺寸的新型非线性器件忆阻器拥有断电非易失性等特点,忆阻器的提出为智能信息处理带来了新的解决方案。在非线性电路的构建中加入忆阻器,基于忆阻器的混沌系统拥有更加丰富的动力学行为,产生的混沌信号拥有更佳的伪随机特性,并且在功耗和体积等方面比传统的混沌系统更占优势。使其在图像加密、扩频通讯以及保密通讯等范畴有更高的研究价值。本文研究了忆阻器的数学及物理结构,分析了对应的基本电学特性与非线性特性,结合忆阻器与混沌系统,设计出基于忆阻器的混沌系统及其所对应的电路。紧接着,本文将忆阻器用于人工神经网络,建立了忆阻细胞神经网络。本论文重点探究了4个部分:(1)研究了忆阻元件的数学模型,并建立了其对应的PSPICE模型,利用数值仿真和电路仿真探究其物理机制,进而验证了其电学特性。(2)构造了具有心型吸引子的忆阻混沌系统。不同于以往的忆阻混沌系统,本系统不仅利用了忆阻器的非线性特性,还利用其可调控特性。忆阻器极性的改变会使该系统产生镜像吸引子,并且随着忆阻参数的调整该系统状态能在混沌态、周期态、稳定态之间转换。使得该系统能同时运用到需要产生混沌信号的系统和需要抑制混沌信号的实际运用中。此外,探讨了该忆阻混沌系统的基本特性如Poincaré截面、Lyapunov指数、分岔图等,并建立了对应的PSPICE仿真电路。(3)提出了一个新型忆阻时滞混沌系统。对提出的忆阻时滞混沌系统进行了稳定性分析,确定了显示系统稳定平衡点的相应参数区域。讨论了在不同参数情况下的系统状态,系统呈现出形态各异的混沌吸引子相图,表现出丰富的混沌特性和非线性特性。将忆阻时滞混沌系统用于产生伪随机信号,并经过实验证明所提出的系统具有良好的相关性,同时能获得相对显著的近似熵。该时滞混沌系统具有复杂的动力学行为和良好的随机性,能满足扩频通信和图像加密等众多领域的应用需要。(4)构建出一种新的忆阻细胞神经网络。改进了传统的忆阻突触桥电路,使之除了具有传统突触桥电路的优势外,还具有更加简化的电路和简化的权值变化条件。通过PSPICE仿真模拟了该突触电路能够实现权值运算。另外,将忆阻细胞神经网络用于图像处理的去噪和边缘提取,实验结果表明忆阻细胞神经网络在图像处理的应用中具有良好的效果。所提出的忆阻细胞神经网络可以减小电路尺寸及提高运算速度,电路结构具有更紧凑和更通用的优点,有助于促进人工神经网络的硬件实现。
[Abstract]:With the rapid development of information society, improving the confidentiality, effectiveness and efficiency of information transmission has become the focus of current scientific and technological development. It brings a new solution for intelligent information processing. Adding a resistor to the construction of nonlinear circuit, the chaotic system based on the resistor has more abundant dynamic behavior. The resulting chaotic signals have better pseudorandom characteristics and are superior to the traditional chaotic systems in power consumption and volume. Spread spectrum communication and secure communication have higher research value. In this paper, the mathematical and physical structures of the resistor are studied, and the corresponding basic electrical and nonlinear characteristics are analyzed. The chaotic system based on the resistive device and its corresponding circuit are designed. Then, in this paper, the resistive device is used in the artificial neural network. In this paper, the mathematical model of the memory element is studied, and the corresponding PSPICE model is established. The physical mechanism of the memory element is explored by numerical simulation and circuit simulation. Furthermore, the electrical properties of this system are verified. (2) A kind of amnesia chaotic system with heart attractor is constructed. Different from the previous amnesia chaotic system, this system not only utilizes the nonlinear characteristics of the amnesia, but also makes use of the nonlinear characteristics of the mnemonic system. The change of the polarity of the resistor will cause the system to produce a mirror attractor, and the state of the system can be in the chaotic state, the periodic state with the adjustment of the parameters of the amnesia. The system can be applied to both the system which needs to produce chaotic signal and the practical application of chaotic signal suppression. In addition, the basic characteristics of the system such as Poincar 茅 section Lyapunov exponent, bifurcation diagram and so on are discussed. A new type of chaotic system with memory delay is proposed, and the stability of the proposed chaotic system is analyzed. The corresponding parameter regions of the stable equilibrium point of the system are determined. The state of the system under different parameters is discussed, and the chaotic attractor phase diagram of different shapes is presented in the system. The chaotic system with memory delay is used to generate pseudorandom signals, and the experimental results show that the proposed system has good correlation. At the same time, a relatively significant approximate entropy can be obtained. The chaotic system with time delay has complex dynamic behavior and good randomness. It can meet the needs of many applications, such as spread spectrum communication and image encryption, etc.) A new amnesia cell neural network is constructed, which improves the traditional mnemonic bridge circuit and makes it have the advantage of the traditional synaptic bridge circuit. The synaptic circuit is simulated by PSPICE to realize weight operation. In addition, the memory cell neural network is used for image processing denoising and edge extraction. The experimental results show that the memory cell neural network has a good effect in image processing. The proposed neural network can reduce the size of the circuit and increase the speed of operation. The circuit structure has the advantages of more compact and more general. It is helpful to promote the hardware implementation of artificial neural network.
【学位授予单位】：西南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN60;O415.5

【参考文献】