基于忆容器件的神经形态计算研究进展

人工智能的快速发展需要人工智能专用硬件的快速发展,受人脑存算一体、并行处理启发而构建的包含突触与神经元的神经形态计算架构,可以有效地降低人工智能中计算工作的能耗.记忆元件在神经形态计算的硬件实现中展现出巨大的应用价值;相比传统器件,用忆阻器构建突触、神经元能极大地降低计算能耗,然而在基于忆阻器构建的神经网络中,更新、读取等操作存在由忆阻电压电流造成的系统性能量损失.忆容器作为忆阻器衍生器件,被认为是实现低耗能神经网络的潜在器件,引起国内外研究者关注.本文综述了实物/仿真忆容器件及其在神经形态计算中的最新进展,主要包括目:前实物/仿真忆容器原理与特性,代表性的忆容突触、神经元及神经形态计算架构,并通过总结近年来忆容器研究所取得的成果,对当前该领域面临的挑战及未来忆容神经网络发展的重点进行总结与展望.     

A mathematical analysis: From memristor to fracmemristor

The memristor is also a basic electronic component, just like resistors, capacitors and inductors. It is a nonlinear device with memory characteristics. In 2008, with HP's announcement of the discovery of the TiO₂ memristor, the new memristor system, memory capacitor (memcapacitor) and memory inductor (meminductor) were derived. Fractional-order calculus has the characteristics of non-locality, weak singularity and long term memory which traditional integer-order calculus does not have, and can accurately portray or model real-world problems better than the classic integer-order calculus. In recent years, researchers have extended the modeling method of memristor by fractional calculus, and proposed the fractional-order memristor, but its concept is not unified. This paper reviews the existing memristive elements, including integer-order memristor systems and fractional-order memristor systems. We analyze their similarities and differences, give the derivation process, circuit schematic diagrams, and an outlook on the development direction of fractional-order memristive elements.     

Investigation on the dynamic behaviors of the coupled memcapacitor-based circuits

In this paper, by referring to the concept of coupled memristors(MRs) and considering the flux coupling connection,the constitutive relations for describing the coupled memcapacitors(MCs) are theoretically deduced. The dynamic behaviors of dual coupled MCs in serial and parallel connections are analyzed in terms of identical or opposite polarities for the first time. Based on the derived constitutive relations of the two coupled MCs, the modified relaxation oscillators(ROs)are obtained with the purpose of achieving controllable oscillation frequency and duty cycle. In consideration of different parameter configurations, the experimental investigation is carried out by using practical off-the-shelf circuit components to verify the correction of the theoretical calculation with numerical simulation of the coupled MCs and its application in ROs.     

A Caputo–Fabrizio Fractional-Order Model of HIV/AIDS with a Treatment Compartment: Sensitivity Analysis and Optimal Control Strategies

Although most of the early research studies on fractional-order systems were based on the Caputo or Riemann–Liouville fractional-order derivatives, it has recently been proven that these methods have some drawbacks. For instance, kernels of these methods have a singularity that occurs at the endpoint of an interval of definition. Thus, to overcome this issue, several new definitions of fractional derivatives have been introduced. The Caputo–Fabrizio fractional order is one of these nonsingular definitions. This paper is concerned with the analyses and design of an optimal control strategy for a Caputo–Fabrizio fractional-order model of the HIV/AIDS epidemic. The Caputo–Fabrizio fractional-order model of HIV/AIDS is considered to prevent the singularity problem, which is a real concern in the modeling of real-world systems and phenomena. Firstly, in order to find out how the population of each compartment can be controlled, sensitivity analyses were conducted. Based on the sensitivity analyses, the most effective agents in disease transmission and prevalence were selected as control inputs. In this way, a modified Caputo–Fabrizio fractional-order model of the HIV/AIDS epidemic is proposed. By changing the contact rate of susceptible and infectious people, the atraumatic restorative treatment rate of the treated compartment individuals, and the sexual habits of susceptible people, optimal control was designed. Lastly, simulation results that demonstrate the appropriate performance of the Caputo–Fabrizio fractional-order model and proposed control scheme are illustrated.     

负参数空间分数阶Chua系统的动力学行为及实验验证

Chua系统展现出丰富的动力学行为,易于电路实现,因而成为混沌研究的经典范例.然而,现有针对Chua系统的研究大都局限于系统的正参数空间.基于分数阶的时域求解法,研究了分数阶Chua系统在负参数空间下的动力学行为.采用分数阶稳定性理论分析了系统平衡点的稳定性,用分岔图、最大李雅普诺夫指数研究了系统控制参数和阶次变化时系统的动力学行为.为了实验验证系统的动力学行为,采用运放、电阻、电容等模拟器件实现了负参数空间下的分数阶Chua系统,实验结果与数值仿真结果完全一致.该研究成果对进一步完善Chua系统,推动Chua系统在混沌中的应用具有参考价值.     

He’s Homotopy Perturbation Method and Fractional Complex Transform for Analysis Time Fractional Fornberg-Whitham Equation

In this article, time fractional Fornberg-Whitham equation of He’s fractional derivative is studied. To transform the fractional model into its equivalent differential equation, the fractional complex transform is used and He’s homotopy perturbation method is implemented to get the approximate analytical solutions of the fractional-order problems. The graphs are plotted to analysis the fractional-order mathematical modeling.     

Accurate capacitance-voltage characterization of organic thin films with current injection

To deal with the invalidation of commonly employed series model and parallel model in capacitance-voltage (C-V) characterization of organic thin films when current injection is significant, a three-element equivalent circuit model is proposed. On this basis, the expression of real capacitance in consideration of current injection is theoretically derived by small-signal analysis method. The validity of the proposed equivalent circuit and theoretical expression are verified by a simulating circuit consisting of a capacitor, a diode, and a resistor. Moreover, the accurate C-V characteristic of an organic thin film device is obtained via theoretical correction of the experimental measuring result, and the real capacitance is 35.7% higher than the directly measured capacitance at 5-V bias in the parallel mode. This work strongly demonstrates the necessity to consider current injection in C-V measurement and provides a strategy for accurate C-V characterization experimentally.     

基于状态观测器的分数阶时滞混沌系统同步研究

研究分数阶时滞混沌系统同步问题,基于状态观测器方法和分数阶系统稳定性理论,设计分数阶时滞混沌系统同步控制器,使得分数阶时滞混沌系统达到同步,同时给出了数学证明过程.该同步控制器采用驱动系统和响应系统的输出变量进行设计,无需驱动系统和响应系统的状态变量,简化了控制器的设计,提高了控制器的实用性.利用Lyapunov稳定性理论和分数阶线性矩阵不等式,研究并给出了同步控制器参数的选择条件.以分数阶时滞Chen混沌系统为例,设计基于状态观测器的同步控制器,实现了分数阶时滞Chen混沌系统同步,并将其应用于保密通信系统中.仿真结果证明了该同步方法的有效性.     

含分数阶微分的线性单自由度振子的动力学分析

研究了一个含分数阶微分的线性单自由度振子, 通过平均法得到了系统的近似解析解. 在近似解中, 分数阶微分项的系数和阶次以等效线性阻尼和等效线性刚度的形式影响着系统的动力学特性, 这一点与现有文献中直接将分数阶微分项归类为阻尼进行处理的方法完全不同. 比较了近似解析解和数值解, 二者的符合精度很高, 证明了近似解析解的准确性. 分析了分数阶系数和分数阶阶次对系统响应特性的影响, 发现分数阶系数和分数阶阶次都既可以通过等效线性阻尼影响系统的共振振幅, 又可以通过等效线性刚度影响系统的共振频率.     

Analysis of Connection Availability in an All-Optical Mesh Network

This paper presents the method of an availability performance analysis that can be used for all-optical networks. We compare different schemes of protection and restoration in a fully connected mesh network based on optical cross-connects, optical fibers, and fiber amplifiers. Specifically treated is an optical mesh network physically mapped onto a Scandinavian backbone. In order to compare different survivability strategies, we selected two nodes located roughly 420 km apart. The goal of this work is to point out possible ways to improve connection availability. A number of survivability schemes are compared, for which the availability model of logical connection is derived and a calculation procedure is described. Our results show that unprotected schemes cannot ascertain reasonably high requirements on connection availability in the backbone network. Thus, if a redundant structure of logical connection is considered, the particular survivability scheme should be evaluated in order to maximize connection availability.     

Design of a memcapacitor emulator based on a memristor

Since Hewlett-Packard?s solid state implementation of a memristive system in 2008, a lot of research has begun to develop applications using this new component. As it is not available as a two terminal device yet, emulating circuits must be developed. Building on our previous work to develop an analog model of a memristor, a memcapacitor emulator is proposed in this Letter. This model can be realized by transforming a memristor emulator to a memcapacitor emulator. The characteristics of a memcapacitor are based on the theory proposed by L.O. Chua. The transformation process is described in detail in this Letter. Simulation and experimental results are given as well.     

忆阻器、忆容器和忆感器的Simulink建模及其特性分析

忆阻器、忆容器和忆感器均是具有记忆特性的新型非线性电路元件,也被称为记忆元件.以三种电路元件的通用数学模型为依据,从数学分析的角度,对忆阻器、忆容器和忆感器的Simulink模型进行了建立.在Simulink模型中体现了记忆元件对历史状态和系统状态变量的依赖性,正确表现出其独特的记忆特性.通过一系列仿真分析,得到了忆阻器、忆容器和忆感器的元件特性,验证了模型的有效性.此外,通过对三者在不同参数、不同激励下的电路特性分析,得到了三种记忆元件等效模型随频率和幅值变化的规律,为以后忆阻器、忆容器和忆感器基于Simulink的仿真研究和应用研究奠定基础.     

Firing multistability,symmetry, bubbles of a Shinriki oscillator with mem-elements

In this paper, a novel memristive Shinriki oscillator is established by adding a flux-controlled memristor and a charge-controlled memcapacitor. The similar symmetry is observed with the Biolek memcapacitor model. Bubbles are also investigated with varying system parameters and initial conditions from the perspectives of cascade and location. Extreme multistability occurring with different initial conditions is clearly found in attraction basins. Different firing patterns with coexisting spiking, bursting and chaotic patterns are discussed in Shinriki oscillator. Spiking and chaotic firing with periodic coexisting large limit cycle and quiescent state are detected. In this oscillator, circle/circle buster is also observed, whose quiescent state is a small amplitude oscillation and spiking state has multiple peaks. The firing multistability of the adjusted Shinriki oscillator is implemented by a Field Programmable Gate Array (FPGA). The experimental results validate the accuracy of the numerical analysis, which shows the proposed oscillator is good for the relevant research and application.     

聚谷氨酸发酵过程中ATR-FTIR光谱信号的分数阶基线校正

采用衰减全反射傅里叶变换红外光谱法（ATR-FTIR）,结合多元校正模型对γ-聚谷氨酸（γ-PGA）发酵过程中两种主要底物葡萄糖和谷氨酸钠的浓度进行间接测量,为优化发酵系统控制提供重要的反馈信息。光谱测量中经常出现的基线漂移会严重影响后续多元校正模型的性能,需要采用基线校正算法对光谱进行预处理。现有流行的基线校正算法多数是基于Whittaker Smoother（WS）平滑算法,这些算法均采用整数阶微分对拟合基线进行约束,表达能力有限。针对现有基线校正算法中的整数阶微分自适应性差的问题,利用更加灵活的分数阶微分对基线进行约束,提出了一种基于分数阶的基线校正算法,实现对整数阶基线校正的扩展。总共进行了5个批次的γ-PGA发酵实验,并对不同批次和全部批次的ATR-FTIR光谱数据分别进行了分数阶基线校正,模型的预测精度均得到不同程度的提升。实验结果表明,只有在批次2时,基于整数阶的基线校正效果最好;其他批次的基线校正效果最好时的阶次均为分数阶。这也表明了分数阶微分（包含整数阶微分）对基线的约束更加合理。同时发现全部批次的整体基线校正效果远远差于单一批次的效果,原因可能是各批次发酵光谱的基线是不同的,对不同的批次需要选用不同的阶次以获得最佳的基线校正。此外,γ-PGA发酵样品的ATR-FTIR光谱测量是以蒸馏水为背景,会在3 100~3 600 cm-1波数范围内出现负水峰,形成有害的干扰信号;分数阶基线校正后的光谱表明,分数阶基线校正算法将负的水峰当作基线,在一定程度上进行了消除。综上分析,分数阶基线校正算法不仅扩展了传统整数阶基线校正算法的应用范围,也为消除ATR光谱中负的水峰提供了新的解决思路。     

Memcapacitor model and its application in a chaotic oscillator

A memcapacitor is a new type of memory capacitor. Before the advent of practical memcapacitor, the prospective studies on its models and potential applications are of importance. For this purpose, we establish a mathematical memcapacitor model and a corresponding circuit model. As a potential application, based on the model, a memcapacitor oscillator is designed, with its basic dynamic characteristics analyzed theoretically and experimentally. Some circuit variables such as charge, flux, and integral of charge, which are difficult to measure, are observed and measured via simulations and experiments. Analysis results show that besides the typical period-doubling bifurcations and period-3 windows, sustained chaos with constant Lyapunov exponents occurs. Moreover, this oscillator also exhibits abrupt chaos and some novel bifurcations.In addition, based on the digital signal processing(DSP) technology, a scheme of digitally realizing this memcapacitor oscillator is provided. Then the statistical properties of the chaotic sequences generated from the oscillator are tested by using the test suit of the National Institute of Standards and Technology(NIST). The tested randomness definitely reaches the standards of NIST, and is better than that of the well-known Lorenz system.     

Analysis and implementation of new fractional-order multi-scroll hidden attractors

To improve the complexity of chaotic signals,in this paper we first put forward a new three-dimensional quadratic fractional-order multi-scroll hidden chaotic system,then we use the Adomian decomposition algorithm to solve the proposed fractional-order chaotic system and obtain the chaotic phase diagrams of different orders,as well as the Lyaponov exponent spectrum,bifurcation diagram,and SE complexity of the 0.99-order system.In the process of analyzing the system,we find that the system possesses the dynamic behaviors of hidden attractors and hidden bifurcations.Next,we also propose a method of using the Lyapunov exponents to describe the basins of attraction of the chaotic system in the matlab environment for the first time,and obtain the basins of attraction under different order conditions.Finally,we construct an analog circuit system of the fractional-order chaotic system by using an equivalent circuit module of the fractional-order integral operators,thus realizing the 0.9-order multi-scroll hidden chaotic attractors.     

Firing activities in a fractional-order Hindmarsh-Rose neuron with multistable memristor as autapse

Considering the fact that memristors have the characteristics similar to biological synapses, a fractional-order multistable memristor is proposed in this paper. It is verified that the fractional-order memristor has multiple local active regions and multiple stable hysteresis loops, and the influence of fractional-order on its nonvolatility is also revealed. Then by considering the fractional-order memristor as an autapse of Hindmarsh-Rose (HR) neuron model, a fractional-order memristive neuron model is developed. The effects of the initial value, external excitation current, coupling strength and fractional-order on the firing behavior are discussed by time series, phase diagram, Lyapunov exponent and inter spike interval (ISI) bifurcation diagram. Three coexisting firing patterns, including irregular asymptotically periodic (A-periodic) bursting, A-periodic bursting and chaotic bursting, dependent on the memristor initial values, are observed. It is also revealed that the fractional-order can not only induce the transition of firing patterns, but also change the firing frequency of the neuron. Finally, a neuron circuit with variable fractional-order is designed to verify the numerical simulations.     

纳米孔隙薄膜中孔隙率与折射率关系的研究

用时域有限差分算法(FDTD)构造了纳米孔隙高分子薄膜的网格结构，并模拟了理想单色平面光波在纳米孔隙高分子薄膜中的传输过程，计算出不同孔隙率的薄膜所对应的等效折射率。模拟结果显示纳米孔隙薄膜的等效折射率与孔隙率呈线性递减的关系。还利用固体材料学中的混合材料等效介电常量理论建立了一种混合介质并联理论模型，很好地解释了纳米孔隙高分子薄膜孔隙率与等效折射率的关系而且得到了纳米孔隙高分子薄膜等效折射率与孔隙率的函数关系式。将模拟结果与混合介质并联理论模型相比较，发现该理论模型的函数曲线与时域有限差分算法模拟的结果相互吻合。     

Experimental characterization and modeling of microsliding on a small cantilever quartz beam

The design of mechanical systems requires various studies in order to ensure an optimal behavior during operation. In particular, the study of its dynamic behavior makes it possible to evaluate the role of a connection in the energy dissipation mechanisms. In this context, an experimental setup dedicated to small structures has been developed to quantify damping due to microsliding at the beam–clamp interface. The mechanical characterization of the clamped connection is carried out by experimental dynamic tests on a free-clamped structure. The instantaneous frequencies and damping are identified by the wavelet transform technique of a slightly nonlinear system. In parallel, numerical prediction of the equivalent damping is achieved thanks to the implementation of the regularized Coulomb law in a finite element model. A genetic algorithm and artificial neural networks are used to update the stiffness parameter and the friction coefficient. The optimized model is in good agreement with experimental results. It allows for determining the spatial distribution of microsliding and tangential force along the contact interface. The dissipated energy and equivalent damping are finally deduced according to the dynamic deflection of the free part of the beam.