基于主成分分析与BP神经网络的识别方法研究

利用BP神经网络对红外目标进行识别之前,若不对原始样本数据进行预处理与特征提取,一方面使识别结果准确性降低,另一方面使BP神经网络的结构复杂化,采用主成分分析法可解决这些问题。主成分分析法能较好地提取表征样本的少数几个主分量,由该方法的特点可知,这几个主分量彼此不相关,非常符合特征优化的要求。研究结果表明,用该方法处理后的结果数据输入BP神经网络．提高了识别正确率,减少了训练时间,同时也简化了网络结构。将两种常见的模式识别方法结合用于红外目标识别：先由主成分分析法对原始样本数据进行精简处理,然后再由BP神经网络法进行分类识别,与传统的单一识别方法相比,准确度得到提高,计算量大为减少。     

Privileged Treatment of UMTS Subscribers in GSM Networks

Different user segments have various requirements and expectations towards the performance of mobile networks. Subscribers having experienced the high quality of UMTS networks desire to maintain high speech quality and excellent data throughput also in areas of missing UMTS but existing GSM coverage. In GSM networks a privileged treatment of UMTS subscribers by means of proper resource allocation provides a substantial quality improvement with respect to standard GSM subscribers. This strategy allows network operators to reduce the performance gap between both network areas experienced by UMTS subscribers. A detailed study on the performance of circuit switched speech and packet data services has been performed based on system level simulations. The results show significant speech quality advantages for users with dual-RAT terminals compared to standard GSM users as well as notably higher data throughput rates.     

Modeling and implementation of 5G edge caching over satellite

The fifth‐generation (5G) wireless networks have to deal with the high data rate and stringent latency requirements due to the massive invasion of connected devices and data‐hungry applications. Edge caching is a promising technique to overcome these challenges by prefetching the content closer to the end users at the edge node's local storage. In this paper, we analyze the performance of edge caching 5G networks with the aid of satellite communication systems. First, we investigate the satellite‐aided edge caching systems in two promising use cases: (a) in dense urban areas and (b) in sparsely populated regions, eg, rural areas. Second, we study the effectiveness of satellite systems via the proposed satellite‐aided caching algorithm, which can be used in three configurations: (a) mono‐beam satellite, (b) multi‐beam satellite, and (c) hybrid mode. Third, the proposed caching algorithm is evaluated by using both empirical Zipf‐distribution data and the more realistic Movielens dataset. Last but not least, the proposed caching scheme is implemented and tested by our developed demonstrators which allow real‐time analysis of the cache hit ratio and cost analysis.     

基于小波神经网络的氧气顶回转炉口火焰温度多光谱测量

氧气顶回转炉(BOF)口火焰温度的分布是对炉内钢水温度和成分含量判定的一项重要依据。通过对炉口350~1100 nm光谱数据的分析,炉口火焰光谱为“带状”辐射重叠在连续的或“黑体”辐射上,在可见光波段有明显的辐射能力。以在南京钢铁公司炼钢炉前在线采集的400炉光谱数据为研究对象,应用小波分析和神经网络的两大类模型交叉结合的方式对炉口火焰温度进行建模预测,并对预测结果做出分析。结果表明,紧致型小波神经网络在预测中取得更佳的效果,基于多光谱测温理论的小波神经网络预测的结果与副枪测量的温度误差能够在理想的范围内。     

Perturbation analysis: the state of the art and research issuesexplained via the GI/G/1 queue

Perturbation analysis (PA) of discrete event dynamic systems (DEDS) enables parameter sensitivities of DEDS to be obtained by observing a single sample path of the system. A simple GI/G/1 system is used to give an introduction to PA and illustrate the basic theoretical issues involved in this technique. The application of PA to networks of queues is covered briefly, and then some of the recent extensions to PA are discussed. It is shown that many interesting open questions remain for PA, and areas for research are indicated     

Clock distribution networks in synchronous digital integratedcircuits

Clock distribution networks synchronize the flow of data signals among synchronous data paths. The design of these networks can dramatically affect system-wide performance and reliability. A theoretical background of clock skew is provided in order to better understand how clock distribution networks interact with data paths. Minimum and maximum timing constraints are developed from the relative timing between the localized clock skew and the data paths. These constraint relationships are reviewed, and compensating design techniques are discussed. The field of clock distribution network design and analysis can be grouped into a number of subtopics: 1) circuit and layout techniques for structured custom digital integrated circuits; 2) the automated layout and synthesis of clock distribution networks with application to automated placement and routing of gate arrays, standard cells and larger block-oriented circuits; 3) the analysis and modeling of the timing characteristics of clock distribution networks; and 4) the scheduling of the optimal timing characteristics of clock distribution networks based on architectural and functional performance requirements. Each of these areas is described the clock distribution networks of specific industrial circuits are surveyed and future trends are discussed     

无卷边单分子层WS2纳米片原子结构的高分辨电子显微术研究

本文利用高分辨电子显微术(HRTEM)确定了无卷边的单分子层WS2纳米片的原子结构像.通过对样品不同区域HRTEM像进行傅里叶变换(FFT)分析,得到了各个区域的WS2纳米片的晶体取向,确定了WS2纳米片分子层数分布.根据电子显微像强度与样品层数的线性关系以及WS2纳米片的单分子层数分布模型,确定不同区域WS2纳米片的分子层数.为了进一步定量分析结果的正确性,通过模拟电子显微像在不同碳膜厚度及不同成像条件下WS2纳米片的HRTEM像,也确定了在像衬度上与实验像很好的匹配.     

Small‐cell intensities for cellular network affected by cell‐edge and hotspot

At present, every network operator is looking for a cost‐effective and spectrally efficient solution not only for better coverage but also to offer higher data rates. Stochastic‐based heterogeneous cellular networks have taken over the thoroughly planned traditional hexagonal grid networks. Heterogeneous networks typically include multiple tiers of base stations and small cells. This paper presents an analysis of a cellular network with cell‐edge, uncovered, and hotspot (HS) users based on spatial bivariate Poisson Point Process. The work compares the relationship between small cell intensity for cell‐edge and HS user's intensity within macrocell area. The variations in small cell intensity for uncovered and HS users outside macrocells are also presented. The simulation results depict higher small cell intensity in cell‐edge areas as compared with HS areas within the macrocells, while for outside of macrocells, small cell intensity in HS areas is much higher than that for uncovered areas. The expected numbers of small cell inside and outside macrocells are obtained along with the total new cells required for the proposed model.     

Authentication and Billing Protocols for the Integration of WLAN and 3G Networks

Wireless communications have developed rapidly and have been applied for many services. Cellular (the third-generation) mobile networks and wireless local area network (WLAN) are two important technologies for providing wireless communications. The third-generation (3G) networks provide wider service areas, and “always-o” and ubiquitous connectivity with low-speed data rate. WLAN networks offer higher data rate and the easy compatibility of wired Internet, but cover smaller areas. In fact, 3G and WLAN possess complementary properties. Integrating 3G and WLAN networks may offer subscribers high-speed wireless data services and ubiquitous connectivity. For integrating two heterogeneous networks, several issues should be involved, authentication, billing, quality of service, and seamless roaming between 3G and WLAN networks. In this paper, we address the authentication and billing problems and propose two protocols that provide both authentication and billing services. One protocol utilizes a one-time password approach to authenticate subscribers. This protocol is efficient in both computation time and authentication procedures. Because of the restrictions of the password-based approach, this protocol could not offer the non-repudiation property for the billing problem. Another protocol is constructed on a public-key-based system (i.e., certificates). Although it requires more computation time than the password-based approach, non-repudiation is guaranteed. Performance analysis simulation results are given to validate our two protocols.     

一种基于帧图像的动态视觉传感器样本集建模方法

动态视觉传感器（Dynamic Vision Sensor,DVS）利用事件驱动原理实现运动目标的快速提取,具有低延迟、低存储空间和高动态范围等优势.目前研究表明,基于DVS的神经网络在目标检测等领域具有明显的速度优势.但是,这类神经网络在训练时所需要的样本集主要依赖DVS相机产生,缺少高效的样本集生成方法,这制约了这类神经网络的应用与发展.本文根据DVS原理,提出了一种基于帧图像的DVS建模以及样本集建模方法.该方法设定每个像素单元独立工作,经过动态差分和逻辑判断后输出触发的地址-事件数据,这些数据通过编码和归一化生成神经网络训练时所需要的样本集.通过对MNIST和CIFAR-10样本集建模的实验结果表明,该建模方法效果与DVS相机基本一致;与基于帧图像的存储方式相比,该样本集可以明显减少存储空间.该方法所生成样本集已经通过卷积神经网络训练和测试验证.     

UMTS-WLAN异构网络互联的方案研究

通用移动通信系统(UMTS)为移动数据用户提供了较广阔的覆盖范围,而无线局域网(WLAN)能在局部热点地区提供较高的接入带宽.两种网络的显著优势能够相互结合,为处于异构网络覆盖地区的用户提供网间无缝连接.本文总结了目前UMTS-WLAN互联的三种实施方案:移动IP、互联网关和仿真器.在分别介绍三种方案的体系架构、互联机制和切换流程后从互联层次、信令和数据流向、对现有网络的改造、安全认证机制及用户管理方式等不同层面分析三种方案的特点.     

A knowledge-based system for highway network extraction

A knowledge-based system has been set up to extract highway networks automatically from Landsat thematic mapper (TM) imagery. The system was developed using image processing algorithms and some artificial intelligence techniques. It inputs raw TM data and outputs an interpreted image clearly identifying highway and road networks. The main algorithms in the system are discussed in detail, results are presented, and features of the system are examined. Landsat TM images of study areas in Toronto, Canada, are presented by way of illustration     

An algorithmic solution for computing circle intersection areas and its applications to wireless communications

A novel iterative algorithm for the efficient computation of the intersection areas of an arbitrary number of circles is presented. The algorithm, on the basis of a trellis structure, hinges on two geometric results, which allow the existence check and the computation of the area of the intersection regions generated by more than three circles by simple algebraic manipulations of the intersection areas of a smaller number of circles. The presented algorithm is a powerful tool for the performance analysis of wireless networks and finds many applications, ranging from sensor to cellular networks. As an example of practical application, an insightful study of the uplink outage probability in a wireless network with cooperative access points as a function of the transmission power and access point density is presented. Copyright © 2012 John Wiley & Sons, Ltd.     

战术自组网分层安全框架研究

未来战术通信网络将使用VHF／UHF无线电进行语音和数据通信,战术通信网络相对固网而言,具有低带宽、高误码率和节点移动性的特点。由于组网对战术通信而言相对较新,其安全性尚未得到完备的考虑,尽管移动Adhoc网络（MANETs）的安全威胁已经被广泛研究,但是不同通信手段具有不同的特点,其安全威胁不一定类似。文中通过分析已有的MANET安全威胁提出了一种新的方法,以评估其对战术网络的潜在影响。尽管这些分析并不覆盖所有可能的威胁,但它具有优势并且能定位主要问题。基于这个分析,提出一个分层的服务框架并集成各层的安全功能。     

Information-intensive wireless sensor networks: potential and challenges

Conventional wireless sensor networks rely mostly on simple scalar data (such as temperature or humidity) and specialize in single-purpose applications. Taking a fundamental departure, in this article we motivate information-rich wireless video sensor networks that emulate the compound eyes found in certain arthropods. Although constrained by scarce resources, sensor nodes can only serve extremely low-resolution video streams; the availability of vast amount of such streams due to deployment redundance can suffice for the need of information hungry applications. Unfortunately, the unique characteristics of wireless video sensor networks will introduce novel uncertainty-driven challenges in the information-intensive and yet resource-constrained environment. Correspondingly, we describe key research problems in the areas of networking, security, sensor design, and video-data analysis     

基于人工智能的跨媒体感知与分析技术研究

随着新媒体技术的飞速发展,传统方法已难以准确表达具有人工智能属性的复杂知识结构,跨媒体成为大家关注的焦点。当前,媒体数据感知与分析已经从文本、语音、图像以及视频等单一媒体形态向覆盖网络空间与物理空间的跨媒体融合转变。研究满足新一代人工智能发展规划的跨媒体感知和分析技术体系,并依托知识图谱、长短时记忆网络以及卷积神经网络等技术,实现多通道网络数据爬取、实体统一表征、文本语义识别以及视图像分类等,可有效支撑舆情分析、新闻追踪以及情报获取等领域的跨媒体应用。     

基于极性复指数变换的图像伪造检测算法

为解决当前图像伪造检测方法在识别复制内容区域时忽略了颜色信息和不同颜色分量之间的相关性,使其对伪造内容的定位与检测准确度不理想的问题,设计了基于改进的加速稳健特征(SURF)描述符与多元极性复指数变换的图像伪造检测算法。引入高斯低通滤波器,对彩色图像完成过滤,以消除噪声,再计算滤波图像的颜色不变性,用其替代SURF描述符中的灰度分量,对SURF方法予以改进,获取新的Hessian矩阵,充分检测彩色图像中的兴趣点;随后,利用这些兴趣点来构建一组连通的Delaunay三角网。基于四元极性复指数变换,充分考虑不同颜色分量之间的相关性,有效提取三角网的局部视觉特征;计算视觉特征之间的欧式距离,根据预设阈值,对三角网实施配准;最后,引入随机样本一致性,剔除错误匹配的三角网,并定义后处理方法,检测出复制伪造区域。测试数据显示：相对已有的复制-粘贴伪造检测方法,在多种几何变换条件下,所提方法具有更高的伪造检测准确性。     

Macrocell path loss prediction using artificial intelligence techniques

The prediction of propagation loss is a practical non-linear function approximation problem which linear regression or auto-regression models are limited in their ability to handle. However, some computational Intelligence techniques such as artificial neural networks (ANNs) and adaptive neuro-fuzzy inference systems (ANFISs) have been shown to have great ability to handle non-linear function approximation and prediction problems. In this study, the multiple layer perceptron neural network (MLP-NN), radial basis function neural network (RBF-NN) and an ANFIS network were trained using actual signal strength measurement taken at certain suburban areas of Bauchi metropolis, Nigeria. The trained networks were then used to predict propagation losses at the stated areas under differing conditions. The predictions were compared with the prediction accuracy of the popular Hata model. It was observed that ANFIS model gave a better fit in all cases having higher R² values in each case and on average is more robust than MLP and RBF models as it generalises better to a different data.     

Impact of Spatial Channel Reuse and Adaptive Modulation on Wireless Mesh Network Capacity

The capacity analysis based on the concept of collision domains may be applied to wireless networks with mesh topology, e.g. wireless backhaul networks for rural areas, sensor networks, etc. The strength of the method is the fact that it may be easily fitted to any MAC protocol, routing algorithm, channel allocation method, and, therefore, is a valuable tool for verification and performance analysis of such algorithms. However, simplifications assumed by some authors may lead to erroneous results, e.g. significantly underestimated capacity of the WMN. Therefore, phenomena specific to wireless networks, such as spatial channel reuse, and features specific to wireless standards, such as link adaptivity, should be included in the analysis, as presented in this paper.     

Research on crowd flows prediction model for 5G demand

The deployment and planning for ultra-dense base stations,multidimensional resource management,and on-off switching in 5G networks rely on the accurate prediction of crowd flows in the specific areas.A deep spatial-temporal network for regional crowd flows prediction was proposed,by using the spatial-temporal data acquired from mobile networks.A deep learning based method was used to model the spatial-temporal dependencies with different scales.External factors were combined further to predict citywide crowd flows.Only data from local regions was applied to model the closeness of properties of the crowd flows,in order to reduce the requirements for transmitting the globe data in real time.It is of importance for improving the performance of 5G networks.The proposed model was evaluated based on call detail record data set.The experiment results show that the proposed model outperforms the other prediction models in term of the prediction precision.