离散Hopfield神经网络和联想记忆的开关电流技术实现

研究了离散Ｈｏｐｆｉｅｌｄ神经网络（ＤＨＮＮ）和联想记忆神经网络的开关电流技术实现，利用多权输入跨导，开关电流延迟器（ＳＩＤ）和可编程电流比较器（ＰＣＣ）实现了离散Ｈｏｐｉｅｌｄ神经网络，并提出了利用离散Ｈｏｐｆｉｅｌｄ神经网络实现自联想记忆时相应的开关电流电路，所提出了开关电流神经网络适宜于超大规模集成，能在低电压（如３．３Ｖ）下工作。     

基于混沌搜索的组合电路测试生成算法

基于组合电路测试生成的离散Hopfield神经网络模型，将混沌搜索与Hopfield网络的梯度算法相结合，利用混沌搜索的内随机性及遍历性来克服梯度算法易于陷于局部极小的缺点，形成一种具有全局搜索能力的测试生成有效算法。该算法综合了随机性和确定性算法的优点，其性能优于一般的随机性算法。实验结果验证了该测试生成算法的有效性。     

基于雪崩晶体管的多电流1550nm激光驱动电路

设计了一种基于雪崩晶体管和可编程逻辑器件的多电流155nm半导体激光驱动电路.该驱动电路包括多种大电窄脉冲生成电路、恒流与保护电路和温控电路三部分.利用51单片机、可编程逻辑器件和雪崩晶体管实现数字式的多种大电流窄脉冲生成电路,易于控制与扩展.恒流与保护电路则借助51单片机的分段程序和高分辨率的数/辛莫转换器实现,精度高.而温控电路采取一种类PWM的控制方式.所设计的驱动电路精度高,实时性好,达到了设计要求.     

一种基于DC/DC变换器的LED驱动电路的设计

提出一种基于电流模式DC/DC变换器的驱动控制电路。该电路可以与恒流电路结合在一起,用作LED驱动。电路由误差放大器、斜坡信号产生电路、电流采样与叠加电路以及PWM比较器四部分构成。采用华虹BCD350工艺进行仿真验证,结果显示,电路成功实现了电流采样信号与斜坡补偿信号的叠加,在Vea信号的控制下,输出了控制功率管关断的PWM脉冲信号。     

基于Hopfield神经网络的中考达线等级预测模型

针对目前中考对青少年的重要性,提出了用学生的平时成绩预测中考达线等级的思想。首先对Hopfield神经网络的结构与联想记忆功能进行了研究,之后建立了基于离散型Hopfield神经网络的预测模型。用MATLAB 7.0对待测样本进行仿真的结果表明,该模型可以有效地对中考达线等级进行预测。     

延迟离散Hopfield网络的稳定性条件

主要利用状态转移方程和定义能量函数的方法对具有时延的离散Hopfield神经网络的动力学行为进行了研究,并且获得了若干稳定性结果.给出了延迟离散Hopfield神经网络收敛于周期为4的极限环的一个充分条件,获得了延迟网络有稳定状态的条件,并且得到了延迟网络既没有稳定状态也没有周期为2的极限环的条件.同时,这些结果用两个例子进行了验证.这些结果推广了一些已有的离散Hopfield神经网络的稳定性结果.     

神经网络的VLSI实现

本文评述了当前神经网络电路实现的关键技术和研究现状,着重讨论了数字、模拟和脉冲流VLSI实现的电路技术及其未来发展。     

通孔最小化的一种神经网络方法

本文在布线的群图模型基础上，利用离散型Hopfield神经网络解决群图的最大割问题，并着重论述了如何跳出局部最优点的问题，从而较好地解决了双层布线通孔最小化问题。算法考虑了许多来自实际的约束，并进行大量的布线实例验证。     

PWM/PFM混合控制DC-DC变换器芯片的设计

结合脉冲宽度调制(PWM)和脉冲频率调制(PFM)功率损耗特点,提出了一种降压型PWM/PFM混合控制DC-DC变换器芯片的电路结构,大大提高了全负载范围转换效率。重点讨论了混合控制策略和PWM/PFM切换电路的设计。Hspice模拟仿真结果验证了设计的正确性。     

复数离散Hopfield网络盲检测64QAM信号

针对复数多电平QAM信号的盲检测问题,该文提出了一个新的复数离散多电平Hopfield神经网络。该网络的实部、虚部各含一个多电平离散激励实函数。该文分析了经典两电平离散Hopfield神经网络能量函数的局限性,构造了一个新的复数多电平神经网的能量函数,并用此能量函数讨论了神经网的稳定性。当该神经网的权矩阵借助接收数据补投影算子构成时,该复数离散多电平Hopfield网络可有效地求解带整数约束的二次规划问题,从而实现QAM信号盲检测。仿真试验表明:该算法所需接收数据较短,就可到达全局真平衡点,计算难度大大降低,具有良好的快速性。     

一种基于小波预处理的模拟电路故障诊断方法

文中提出了一种基于小波预处理的模拟电路故障诊断方法。由于小波分析具有数据压缩和特征提取的特性,我们利用小波变换对电路脉冲信号进行多尺度分解,提取特征向量输入神经网络进行训练。实验表明,该办法可以有效地减少神经网络的训练时间,提高模拟电路故障诊断的准确率。     

Hardware Implementation of a Pulse Density Neural Network Using Simultaneous Perturbation Learning Rule

The choice of the learning scheme is very important in the implementation of neural networks to take advantage of their learning ability. Usually, the back-propagation method is widely used as a learning rule in neural networks. Since back-propagation requires so-called error back propagation to update weights, it is relatively difficult to realize hardware neural networks using the back-propagation method. In this paper, we present a pulse density neural network system with learning ability. As a learning rule, the simultaneous perturbation method is used. The learning rule requires only forward operations of networks to update weights instead of the error back-propaga- tion. Thus, we can construct the network system with learning ability without the need for a complicated circuit that calculates gradients of an error function. Pulse density is used to represent the basic quantities in this system. The pulse system has some attractive properties which includes robustness against a noisy environment. A combina- tion of the simultaneous perturbation learning rule and the pulse density system results in an interesting architec- ture of hardware neural systems. Results for the exclusive OR problem and a simple identity problem are shown.     

数字电路的最优神经网络模型及建立方法

本文研究电路的最优神经网络模型,获得了对任意结构的多输入多输出逻辑电路,都存在一种最优神经网络能表征电路的逻辑功能,通过求解一个线性方程组可以得到这种神经网络的结构.文中也给出了多输入基本门电路的最优神经网络结构及其能量函数的一般表达式.     

模拟数字电路故障诊断新方法

利用小波变换与神经网络相结合的方法,采用能量分布特征提取方法和改进BP算法,给出了一种基于小波变换和BP神经网络相结合的模拟电路故障诊断方法.用正弦信号仿真模拟电路,应用小波变换对模拟电路的采样信号进行多尺度分解,再进行能量分布特征提取,然后利用神经网络对各种状态下的特征向量进行分类识别,实现模拟电路故障诊断.在用神经网络诊断模拟电路的基础上,进行了将神经网络用于数字电路单故障诊断的研究.对两者的实例电路仿真结果表明,神经网络可以有效、方便地实现电路的故障检测和定位,准确率高,为故障诊断的研究提供了一种新思路.     

A dynamic CMOS multiplier for analog VLSI based on exponentialpulse-decay modulation

A clocked, charge-based, CMOS modulator circuit is presented. The circuit, which performs a semilinear multiplication function, has applications in arrayed analog VLSI architectures such as parallel filters and neural network systems. The design presented is simple in structure, uses no operational amplifiers for the actual multiplication function, and uses no power in the static mode. Two-quadrant weighting of an input signal is accomplished by control of the magnitude and decay time of an exponential current pulse, resulting in the delivery of charge packets to a shared capacitive summing bus. The cell is modular in structure and can be fabricated in a standard CMOS process. An analytical derivation of the operation of the circuit, SPICE simulations, and MOSIS fabrication results are presented. The simulation studies indicate that the circuit is inherently tolerant to temperature effects, absolute device sizing errors, and clock-feedthrough transients     

Resolution enhancement in long pulse OTDR for application in structural health monitoring

To improve the range resolution in inexpensive conventional long pulse optical time domain reflectometer (OTDR) for application in structural health monitoring (SHM) and robotic neural network, the Fourier Wavelet Regularized Deconvolution (ForWaRD) based on the adaptive wavelet method is employed. Since Deconvolution is a noise sensitive process, employing the (ForWaRD) method enhances the signal to noise ratio. Simulation for long pulse OTDR system is done and ForWaRD method is employed to improve the resolution of the OTDR system to the order of several centimeters. In this method the resolution is limited by the bandwidth of detector, bandwidth of electronic circuit, and the sampling rate of analog to digital convertor.     

基于空间特征的非线性电路神经网络诊断

本文提出线性模拟电路的单、双、三故障空间特征,采用分段线性模型（PWL）将非线性电路线性化,通过遗传算法求电路的容羞范围,用神经网络对非线性嘲络进行诊断。本文的方法大火减少了模拟计算量,同时,使神经网络的训练过程加快,训练误差减少,并大大提高了诊断的正确率。     

基于MIV和BRBP神经网络的电路板红外诊断方法

针对BP神经网络对于海量数据训练及多维数据训练收敛困难的问题,在使用增加动力项、自适应学习速率等方法的基础上,引入均值影响度算法(MIV)构造了贝叶斯正则化反向传播(BRBP)神经网络,以此提高电子线路板红外故障诊断算法的效率。利用红外测温方式,获取了不同室温及运行状态下电路板中21个元器件温度数据。将此21个参数作为故障诊断模型的初始输入变量,经过MIV算法简约为12个参数输入至BRBP神经网络,进行故障评估和诊断。结果表明:相对于传统的BRBP神经网络,本文设计的基于MIV和BRBP神经网络模型诊断方法极大简化了数据训练的数据量并解决了数据收敛的困难,因此效率更高,用时更省。     

Memristor-based multi-channel pulse coupledneural network for image fusion

Image fusion is widely used in computer vision and image analysis. Considering that the traditional image fusionalgorithm has a certain limitation in multi-channel image fusion, a memristor-based multi-channel pulse coupledneural network (M-MPCNN) for image fusion is proposed. Based on a dual-channel pulse coupled neural network(D-PCNN), a novel multi-channel pulse coupled neural network (M-PCNN) is firstly constructed in this paper.Then the exponential growth dynamic threshold model is used to improve the pulse generation of pulse coupledneural network, which can not only avoid multiple ignitions effectively, but can also improve operational efficiencyand reduce complexity. At the same time, synchronous capture can also enhance image edge, which is moreconducive to image fusion. Finally, the threshold and synaptic characteristics of pulse coupled neural networks(PCNNs) can be well realized by using a memristor-based pulse generator. Experimental results show that theproposed algorithm can fuse multi-source images more effectively than existing state-of-the-art fusion algorithms.     

A neural network modeling approach to circuit optimization andstatistical design

The trend of using accurate models such as physics-based FET models, coupled with the demand for yield optimization results in a computationally challenging task. This paper presents a new approach to microwave circuit optimization and statistical design featuring neural network models at either device or circuit levels. At the device level, the neural network represents a physics-oriented FET model yet without the need to solve device physics equations repeatedly during optimization. At the circuit level, the neural network speeds up optimization by replacing repeated circuit simulations. This method is faster than direct optimization of original device and circuit models. Compared to existing polynomial or table look-up models used in analysis and optimization, the proposed approach has the capability to handle high-dimensional and highly nonlinear problems