Optimized operations by extended X-factor theory including unithours concept

The relationships between tool performance, cycle time, and throughput have been described to achieve the highest productivity. Though tool utilization is low for manually operated tools, long waits prior to product processing lead to a productivity loss in our production line. We focus on operator efficiency in the CMP (chemical mechanical polishing) operation, which, as a typical manually operated tool, influences cycle time. This paper introduces an online/offline unit hour analysis method, which has been designed to optimize operator allocation and headcount in our production line, using X-factor theory in order to achieve the shortest cycle time. By using online/offline unit hour analysis with X-factor theory, the root cause of waiting time can be identified and the relationship between operator headcount and cycle time can be described. The waiting time can be reduced by 35% and the cycle time can be reduced by 25%, according to a simulator with optimized operator allocation and headcount based on online/offline unit hour analysis     

基于数据中心负载分析的自适应延迟调度算法

由于已有的延迟调度算法基于静态的等待时间阈值,因此不能适应云计算数据中心动态的负载变化。针对该问题,提出了等待时间阈值自适应调整模型。基于该模型,设计了自适应延迟调度算法(ADS)。ADS算法通过分析空闲计算节点的到达强度、网络带宽和作业执行状态等参数,自适应调整等待时间阈值,以减少作业响应时间。基于开发的原型系统,验证了自适应调整模型,测试了算法性能。结果表明,ADS算法在作业响应时间等方面优于已有的延迟调度算法。     

基于神经网络的分振幅光偏振仪的数据处理

分振幅光偏振仪(DOAP)是一种高速测量光波偏振态的传感器。提出了一种基于人工神经网络(ANN)的分振幅光偏振仪的数据处理方法,将分振幅光偏振仪中电路系统输出的电信号作为神经网络的输入,入射光的斯托克斯参数作为神经网络的输出,建立一个前向多层神经网络模型。通过网络训练,使该网络确立了电路系统输出电信号与入射光斯托克斯参数之间的映射关系。由测量时得到的电信号,利用训练后的神经网络可以计算出待测的入射光的斯托克斯参数。测试结果表明,在测量精度方面,该方法获得的斯托克斯参数的总均方根偏差为1.9%,略优于基于矩阵运算的数据处理方法。     

高光谱成像技术检测冷却羊肉表面细菌总数

为了对冷却羊肉表面细菌总数进行无损检测,采用不同波段范围高光谱成像系统结合多种建模方法建立预测模型,进行理论分析和实验验证。分别在400nm~110nm和900nm~1700nm波长范围内获取冷却羊肉样本的高光谱图像信息,结合偏最小二乘和人工神经网络(反向人工神经网络和径向基人工神经网络)建立预测模型。结果表明,神经网络建模效果优于偏最小二乘;其中,径向基人工神经网络模型在400nm~1100nm和900nm~1700nm波长范围内相关系数分别为0.9872和0.9988,均方根误差分别为0.8210和0.2507,预测效果最好;而900nm~1700nm波长范围为最佳建模波长。这一结果说明利用高光谱图像技术对冷却羊肉表面细菌总数进行快速无损检测是可行的。     

BP网络的Matlab实现及应用研究

人工神经网络以其具有信息的分布存储、并行处理以及自学习能力等优点,已经在信息处理、模式识别、智能控制及系统建模等领域得到越来越广泛的应用。他的基于误差反向传播算法的多层前馈网络,即BP网络在非线性建模、函数逼近和模式识别中有广泛的应用,介绍了BP网络的基本原理,分析了Matlab人工神经网络工具箱中有关BP网络的工具函数,并给出了部分重要工具函数的实际应用。     

Receiver Design for Uplink Multiuser Code Division Multiple Access Communication System Based on Neural Network

In this paper, we consider the receiver design problem for the uplink multiuser code division multiple access (CDMA) communication system based on the neural network technique. The uplink multiuser CDMA communication system model is described in the form of space–time domain through antenna array and multipath fading expression. Novel suitable neural network technique is proposed as an effective signal processing method for the receiver of such an uplink multiuser CDMA system. By the appropriate choice of the channel state information for the neural network parameters, the neural network can collectively resolve the effects of both the inter-symbol interference due to the multipath fading channel and the multiple access interference in the receiver of the uplink multiuser CDMA communication system. The dynamics of the proposed neural network receiver for the uplink multiuser CDMA communication system is also studied.     

HTTP malicious traffic detection method based on hybrid structure deep neural network

In response to the HTTP malicious traffic detection problem,a preprocessing method based on cutting mechanism and statistical association was proposed to perform statistical information correlation as well as normalization processing of traffic.Then,a hybrid neural network was proposed based on the combination of raw data and empirical feature engineering.It combined convolutional neural network (CNN) and multilayer perceptron (MLP) to process text and statistical information.The effect of the model was significantly improved compared with traditional machine learning algorithms (e.g.,SVM).The F₁value reached 99.38% and had a lower time complexity.At the same time,a data set consisting of more than 450 000 malicious traffic and more than 20 million non-malicious traffic was created.In addition,prototype system based on model was designed with detection precision of 98.1%~99.99% and recall rate of 97.2%~99.5%.The application is excellent in real network environment.     

Computerized tumor boundary detection using a Hopfield neuralnetwork

The authors present a new approach for detection of brain tumor boundaries in medical images using a Hopfield neural network. The boundary detection problem is formulated as an optimization process that seeks the boundary points to minimize an energy functional based on an active contour model. A modified Hopfield network is constructed to solve the optimization problem. Taking advantage of the collective computational ability and energy convergence capability of the Hopfield network, the authors' method produces the results comparable to those of standard “snakes”-based algorithms, but it requires less computing time. With the parallel processing potential of the Hopfield network, the proposed boundary detection can be implemented for real time processing. Experiments on different magnetic resonance imaging (MRI) data sets show the effectiveness of the authors' approach     

Design of the processing core of a mixed-signal CMOS DTCNN chip for pixel-level snakes

This paper introduces the processing core of a full-custom mixed-signal CMOS chip intended for an active-contour-based technique, the so-called pixel-level snakes (PLS). Among the different parameters to optimize on the top-down design flow our methodology is focused on area. This approach results in a single-instruction-multiple-data chip implemented by a discrete-time cellular neural network with a correspondence between pixel and processing element. This is the first prototype for PLS; an integrated circuit with a 9/spl times/9 resolution manufactured in a 0.25 -/spl mu/m CMOS STMicroelectronics technology process. Awaiting for experimental results, HSPICE simulations prove the validity of the approach introduced here.     

基于卷积神经网络的图像分类算法综述

随着大数据的到来以及计算能力的提高,深度学习(Deep Learning, DL)席卷全球。传统的图像分类方法难以处理庞大的图像数据以及无法满足人们对图像分类精度和速度上的要求,基于卷积神经网络(Convolutional Neural Network, CNN)的图像分类方法冲破了传统图像分类方法的瓶颈,成为目前图像分类的主流算法,如何有效利用卷积神经网络来进行图像分类成为国内外计算机视觉领域研究的热点。本文在对卷积神经网络进行系统的研究并且深入研究卷积神经网络在图像处理中的应用后,给出了基于卷积神经网络的图像分类所采用的主流结构模型、优缺点、时间/空间复杂度、模型训练过程中可能遇到的问题和相应的解决方案,与此同时也对基于深度学习的图像分类拓展模型的生成式对抗网络和胶囊网络进行介绍;然后通过仿真实验验证了在图像分类精度上,基于卷积神经网络的图像分类方法优于传统图像分类方法,同时综合比较了目前较为流行的卷积神经网络模型之间的性能差异并进一步验证了各种模型的优缺点;最后对于过拟合问题、数据集构建方法、生成式对抗网络及胶囊网络性能进行相关实验及分析。      

Widrow-Hoff神经网络学习规则的应用研究

基于线性神经网络原理,提出线性神经网络的模型,并利用Matlab实现Widrow-Hoff神经网络算法。分析Matlab人工神经网络工具箱中有关线性神经网络的工具函数,最后给出线性神经网络在系统辨识中的实际应用。通过对线性神经网络的训练,进一步验证Widrow-Hoff神经网络算法的有效性,以及用其进行系统辨识的高精度拟合性。     

采用深度卷积神经网络方法的毫米波图像目标检测

通过收集大量的毫米波图像并建立相应的人体数据集进行检测,提出基于Faster R-CNN深度学习的方法检测隐藏于人体上的危险物品。该方法将区域建议网络和VGG19训练卷积神经网络模型相结合,构建了面向毫米波图像目标检测的深度卷积神经网络。为了提高毫米波图像的处理能力,采用Caffe深度学习框架在图形处理单元上进行训练和测试。实验结果证明了基于Faster R-CNN深度卷积神经网络的目标检测方法能有效检测毫米波图像中的危险物品,并且目标检测的平均准确率约94%,检测速度约为6 frame/s,对毫米波安检系统的智能化发展有着极其重要的参考价值。     

On the design of Robust‐WDM local area networks

Robust wavelength division multiplexing (Robust‐WDM) is a proposal to realize cost‐effective WDM local area networks (LANs) which can get around the expensive need for laser wavelength stabilization. The type of these networks that relies on an access protocol with aperiodic reservations and lenient‐token‐passing based control channel (the AR/LTP protocol) is promising. We look at the deployment of the AR/LTP analytical model in designing this type of network. The model is used to predict the effect of component and network parameters on the waiting time characteristics of the network. An increase in node operation times (i.e. receiver response time, transmitter select time and reservation period) would result in increasing the average waiting time of a connection request, but the waiting time is more sensitive to the physical span of the network and its size. It is also observed that increasing the inter‐reservation threshold may result in little increase in waiting time up to some limit beyond which delay increases rapidly. Copyright © 2001 John Wiley & Sons, Ltd.     

A High Speed VLSI Architecture for Handwriting Recognition

This article presents PAPRICA-3, a VLSI-oriented architecture for real-time processing of images and its implementation on HACRE, a high-speed, cascadable, 32-processors VLSI slice. The architecture is based on an array of programmable processing elements with the instruction set tailored to image processing, mathematical morphology, and neural networks emulation. Dedicated hardware features allow simultaneous image acquisition, processing, neural network emulation, and a straightforward interface with a hosting PC.HACRE has been fabricated and successfully tested at a clock frequency of 50 MHz. A board hosting up to four chips and providing a 33 MHz PCI interface has been manufactured and used to build BEATR IX, a system for the recognition of handwritten check amounts, by integrating image processing and neural network algorithms (on the board) with context analysis techniques (on the hosting PC).     

一种机场鸟情监测系统

针对民航飞机鸟击事件多发于起飞和着陆阶段的特点,提出了一种基于三坐标搜索雷达、毫米波高分辨观测雷达与光学辅助设备的机场鸟情监测系统。介绍了监测系统的功能、系统组成与工作流程,描述了以神经网络为核心的鸟情信息处理系统以及神经网络的层次结构,探讨了训练模型和工作模型。通过对系统观测数据的特征提取和处理,结合鸟类运动规律,实...     

Neural-network-based delivery time estimates for prioritized 300-mm automatic material handling operations

This paper deals with lot delivery estimates in a 300-mm automatic material handling system (AMHS), which is composed of several intrabay loops. We adopt a neural network approach to estimate the delivery times for both priority and regular lots. A network model is developed for each intrabay loop. Inputs to the proposed neural network model are the combination of transport requirements, automatic material handling resources, and ratios of priority lots against regular ones. A discrete-event simulation model based on the AMHS in a local 300-mm fab is built. Its outputs are adopted for training the neural network model with the back propagation method. The outputs of the neural network model are the expected delivery times of priority and regular lots in the loop, respectively. For a lot to be transported, its expected delivery time along a potential delivery path is estimated by the summation of all the loop delivery times along the path. A shortest path algorithm is used to find the path with the shortest delivery time among all the possible delivery paths. Numerical experiments based on realistic data from a 300-mm fab indicate that this neural network approach is sound and effective for the prediction of average delivery times. Both the delivery times for priority and regular lots get improved. Specially, for the cases of regular lots, our approach dynamically routes the lots according to the traffic conditions so that the potential blockings in busy loops can be avoided. This neural network approach is applicable to implementing a transport time estimator in dynamic lot dispatching and fab scheduling functions in realizing fully automated 300-mm manufacturing.     

Introduction to the modified probabilistic neural network forgeneral signal processing applications

This paper introduces a practical and easy-to-understand network for signal processing called the modified probabilistic neural network (MPNN). It begins with a short introduction to the application of artificial neural networks to signal processing followed by a background and review of the MPNN theory. The MPNN is a regression technique similar to Specht's (1991) general regression neural network, which is based on a single radial basis function kernel whose bandwidth is related to the noise statistics. It has advantages in application to time and spatial series signal processing problems because it is constructed directly and simply from the training signal waveform characteristics or features. An illustrative example involving noisy Doppler-shifted swept frequency sonar signal detection compares the effectiveness of the first- and second-order Volterra, multilayer perceptron neural network, radial basis function neural network, general regression neural network and MPNN filters, demonstrating some features of the MPNN for practical design     

Linear and non‐linear space diversity combining algorithms over fading channels in TDMA systems

This paper investigates the performances of various adaptive algorithms for space diversity combining in time division multiple access (TDMA) digital cellular mobile radio systems. Two linear adaptive algorithms are investigated, the least mean square (LMS) and the square root Kalman (SRK) algorithm. These algorithms are based on the minimization of the mean‐square error. However, the optimal performance can only be obtained using algorithms satisfying the minimum bit error rate (BER) criterion. This criterion can be satisfied using non‐linear signal processing techniques such as artificial neural networks. An artificial neural network combiner model is developed, based on the recurrent neural network (RNN) structure, trained using the real‐time recurrent learning (RTRL) algorithm. It is shown that, for channels characterized by Rician fading, the artificial neural network combiners based on the RNN structure are able to provide significant improvements in the BER performance in comparison with the linear techniques. In particular, improvements are evident in time‐varying channels dominated by inter‐symbol interference. Copyright © 2001 John Wiley & Sons, Ltd.     

用于信号逼近的自适应时延小波神经网络

小波神经网络是一种强有力的函数逼近工具。本文结合时延神经网络和小波分析概念提出一种新的小波神经网络摸型自适应时延小波神经网络(ATDWNN:adaptive timedelay wavelet neural network).ATDWNN可以对同一类存在不同时延的多个信号用同一个超小波(superwavelet)进行逼近。为了训练ATDWNN,本文提出一种基于时间机理的竞争学习算法。实验表明,ATDWNN不仅可以成功地对同一类存在不同时延的多个信号采用同一个超小波进行逼近,而且可以用来估计各样本信号的时延。     

一种改进的用于可靠性分析的神经网络的实现

针对容错系统的可靠性问题,建立基于马尔科夫模型的三层前馈神经网络。提出一种改进的神经网络训练算法,用于包含永久过错．瞬态过错和周期过错影响的一个三模冗余（TMR,Triple Modular Redundancy）系统的可靠性分析。一个全连接的三层神经网络表示一个容错系统的离散时间n状态马尔科夫模型的可靠性。将系统的期望可靠性反馈入网络,当神经网络收敛时,从神经网络的权值中得出设计参数。仿真结果显示,与四层神经网络相比．三层神经网络收敛得更快．收敛可靠十牛更椿沂期望可靠性。