基于MapReduce的交互可视化平台

针对长时间序列的海洋环境信息交互可视化处理问题,本文提出一种基于云环境的交互可视化平台架构.可视化平台以Hadoop为基础,将GPU 、MPI并行计算引入MapReduce处理机制,实现海洋环境空间数据场大规模数据集的检索、抽取、插值计算、特征可视化分析的并行处理,达到海量数据的远程交互可视化处理目的.实验结果表明,本文提出的可视化平台架构可以有效地将MPI、GPU并行计算与MapReduce处理机制相结合,以满足海洋环境信息交互可视化的需求.     

自中心网络生成的高效分布式设计与实现

近年来数据挖掘在电信领域的应用越来越广泛,而自中心网络从人与环境交互的角度为数据分析提出了新的思路,但是由于数据量、数据维度、计算复杂度等原因,传统的计算方式不能应付海量数据的自中心网络生成和分析的需求.本文首先给出了基于mpreduce模型的传统自中心网络生成算法的实现,然后提出了新的基于三角形提取自中心网络生成算法,并给出了基于mapreduce编程模型的实现,该算法针对mapreduce模型和真实的社会网络进行了优化并实现性能提升,最后对两种算法进行了运行时间和10的比较.     

MapReduce框架下的粒概念认知学习系统研究

针对经典的概念学习算法难以处理大规模数据集的问题,本文提出一种基于MapReduce框架的粒概念认知学习并行算法.该算法借鉴认知心理学的知觉和注意认知思想,并融合粒计算的粒转移原理.首先构建适应大数据环境的粒概念并行求解算法,并与经典粒概念构造算法做了对比,在此基础上分别从外延和内涵角度建立了粒概念认知计算系统,然后对给定对象集或属性集进行认知概念学习.实验结果表明,该并行算法是有效的,适合海量数据的粒概念认知学习.     

海量数据信息管理系统设计与实现

为了提高不同场景的绘制效率,对网络中海量数据信息进行可视化管理,设计一种基于三维虚拟的海量数据信息管理系统。系统分为三维图像处理模块、信息感知模块、数据交互模块、三维模型重构渲染模块和输出程序控制模块。采用视景仿真渲染工具Vega Prime进行海量数据信息管理系统的三维立体建模和视景分析,采用3ds MAX软件进行海量数据信息管理建模,结合自适应图形跟踪渲染方法实现网络中海量数据可视化三维图形绘制,提高对海量数据信息的跟踪能力,构建视景分析模型数据库,实现对海量数据信息进行实时管理。仿真结果表明,设计的系统具有很好的三维虚拟重构能力,提高了海量数据显示与交互操作的绘制效率。     

基于Web的空间数据可视化研究

分析了Web环境中空间数据可视化的特征,提出基于结点的面向对象空间数据模型,采用面向对象的方法来表达和管理空间数据;然后设计了一个有效的空间数据可视化体系结构,该体系构架在海量空间数据的组织、管理、传输和显示上有了突破,实现了Web环境下的空间数据可视化.最后给出了该系统的应用实例,该系统具备空间数据的真实感可视化表达、场景实时漫游、交互查询等功能.     

Massive MIMO中通信高效的分布式预编码设计

针对多BBU基带处理架构，提出一种通信高效的分布式预编码方案，旨在降低BBU间前传交互和计算复杂度。首先，提出基于R-WMMSE算法的分布式框架，利用最优解的子空间特性无损压缩交互数据，降低数据交互量。然后设计了2种基于矩阵乘法的可学习压缩模块，通过优化的计算结构和矩阵参数减少参数和计算量，并保持函数表达能力。最后，以可达速率为优化目标，将可学习模块和分布式预编码算法框架联合优化得到最终模型。所提方案可以在更低的数据交互和计算复杂度要求下，实现预编码性能的保障。     

网络环境下高性能计算的可视化研究

网络环境下高性能计算的可视化在宇宙物理学、大数据统计、遥感数据处理等领域发挥重要作用,包括数据管理、数据调度、数据处理与交互等环节,二维可视化技术越来越成熟,目前国内外尚无一种系统平台能够有效的提高数据处理,同步三维显示化问题.     

云环境下一种低成本的数据安全存储和处理框架

提出了一种基于云计算的数据存储和处理框架,支持用户对大规模数据集进行低成本、安全的共享和处理。该框架研究了基于矩阵的数据挖掘算法,重点是安全的Top-k特征向量算法。该算法采用了一种迭代处理模型,授权用户可以与云端展开交互以获得预期结果,交互过程中源矩阵和中间结果均处于保密状态。另外,基于Paillier加密算法和高效的随机向量扰动策略,实现了新算法的安全性。实验结果表明,提出的方法对大型矩阵具有可拓展性,且客户端成本较低。     

Hadoop下基于贝叶斯网络的气象数据挖掘研究

基于朴素贝叶斯的分类器是气象数据挖掘中比较传统的方法,但由于算法要求各属性相互独立,预测精度无法达到要求,且在处理海量数据时算法计算效率受到制约,对此提出一种Hadoop平台下基于离散贝叶斯网络的数据挖掘改进算法。算法不要求属性之间相互独立,且充分结合Hadoop平台适应处理大数据的优点,利用海量数据分析地面气象因素与温度之间的相关性,并由此选取预测因子来训练贝叶斯网络分类器模型,以达到预测温度的目的。实验结果表明,算法不但预测精度明显高于目前短期气候预测中采用的朴素贝叶斯算法,而且极大地提高了运算效率。     

非结构化海量网络数据处理技术研究

为实现网络化测试系统下非结构化海量网络数据的快速分析处理,在关键的算法和系统化集成处理方面提出解决方法。采用内存映射文件方式快速读取海量数据,并设计了时间矩阵算法,用以快速进行同步分析处理;应用分布式中间件方式实现海量数据的并发处理和数据分发,对飞行试验采集的网络数据进行了分析处理,使用这些算法的数据处理软件,可以使处理效率满足飞行试验海量网络数据处理的需求。这些都为新一代机载采集系统应用于飞行试验提供了技术保障。     

信号级协同计算平台架构及应用思考

针对智能化作战对军事电子信息系统计算能力提升的迫切需求,结合云计算的资源虚拟化、大数据的分布式计算等技术,提出了基于嵌入式CPU+ALL（DSP、FPGA、PPC、GPU、AI处理器等）的异构处理的信号级协同计算平台架构,包括弹性、轻量级异构资源虚拟化模型、分布式实时计算框架和智能计算框架等,形成了一套架构统一、资源共用、使用简便的协同计算和智能计算环境。通过战术级无线电认知和智能信号与信息处理两个典型应用场景,探讨了该信号级协同计算平台可能带来的颠覆性效用。     

Microelectronic architectures and devices for signal and symbol processing

Modern microelectronics technology holds the promise of high level, computer-aided design of very complex systems on a single silicon chip. The freedom to create non-standard architectures within this context has stimulated widespread interest in the development of computing structures that offer increased processing speed relative to the von Neumann architecture. This paper describes two architectures that are especially well suited for large scale integration because of their concurrent structure and their use of primarily local data flows. The first architecture is designed to implement the QR matrix decomposition and it can be used to reliably solve the least squares and eigenvalue problems of linear algebra. The second architecture is based on an elementary building block approach to the realization of FIR and IIR lattice digital filters.An especially important issue that is sometimes overlooked in the system level design of an architecture is the impact of the particular scheme for implementing the fundamental operations such as multiplication, rotation, etc., on the performance of the highly parallel computing structure. It is argued that if one can imbed pipelined operations within the concurrent computing structure, then often the resulting system will not only provide substantial processing gain but it can be implemented in such a way that efficient use of chip real estate is achieved.     

Wireless distributed computing in cognitive radio networks

Individual cognitive radio nodes in an ad-hoc cognitive radio network (CRN) have to perform complex data processing operations for several purposes, such as situational awareness and cognitive engine (CE) decision making. In an implementation point of view, each cognitive radio (CR) may not have the computational and power resources to perform these tasks by itself. In this paper, wireless distributed computing (WDC) is presented as a technology that enables multiple resource-constrained nodes to collaborate in computing complex tasks in a distributed manner. This approach has several benefits over the traditional approach of local computing, such as reduced energy and power consumption, reduced burden on the resources of individual nodes, and improved robustness. However, the benefits are negated by the communication overhead involved in WDC. This paper demonstrates the application of WDC to CRNs with the help of an example CE processing task. In addition, the paper analyzes the impact of the wireless environment on WDC scalability in homogeneous and heterogeneous environments. The paper also proposes a workload allocation scheme that utilizes a combination of stochastic optimization and decision-tree search approaches. The results show limitations in the scalability of WDC networks, mainly due to the communication overhead involved in sharing raw data pertaining to delegated computational tasks.     

Hybrid Control and Reservation Architecture for Multidomain Burst Switched Network

Optical burst switching realizes all-optical communication by dividing data transmission system into control plane and data plane. In order to implement control plane two controlling mechanisms such as distributed and centralized control and reservation models have been presented in literature. Both of the architectures have pros and cons associated such as distributed model lacks efficient resource utilization while centralized model is nonscalable. In order to overcome these, a new hybrid control and reservation mechanism has been presented by the authors. The proposed model is based on clustering technique where each cluster consists of cluster head and member nodes. Cluster head performs major functionalities in the network such as path establishment, routing and resource management, fault management, etc., while member nodes perform data switching functionality. In order to implement the architecture, a signaling protocol is also presented. In this paper, we discuss different scenarios required for the implementation of the proposed hybrid model. These scenarios detail the message dialogs exchange between different network nodes and between different software modules during lightpath establishment, data transmission, and connection teardown operations. The goal of efficient resource utilization is achieved through centralized control of network resources while scalability is achieved through dividing the network into multiple domains. Finally, the simulation section shows the mechanism of designing the multidomain architecture, network bandwidth utilization comparison and reduction in end-to-end data transmission latency.      

Memory-efficient high-speed VLSI implementation of multi-level discrete wavelet transform

Memory requirements and critical path are essential for 2-D Discrete Wavelet Transform (DWT). In this paper, we address this problem and develop a memory-efficient high-speed architecture for multi-level two-dimensional DWT. First, dual data scanning technique is first adopted in 2-D 9/7 DWT processing unit to perform lifting operations, which doubles the throughputs per cycle. Second, for 2-D DWT architecture, the proposed Row Transform Unit and Column Transform Unit take advantage of input sample availabilities and provision computing resources accordingly to optimize the processing speed, in which the number of processors is further optimized to significantly reduce the hardware cost. Third, to address the problem of high cost of memory for the immediate computing results from each level and the computation time as resolution level increases, multiple proposed 2-D DWT units were combined to build a parallel multi-level architecture, which can perform up to six levels of 2-D DWT in a resolution level parallel way on any arbitrary image size at competitive hardware cost. Experimental results demonstrated that the proposed scheme achieves improved hardware performance with significantly reduced on-chip memory resource and computational time, which outperforms the-state-of-the-art schemes and makes it desirable in memory-constrained real-time application systems.     

Real-time multimodal medical image processing: a dynamicvolume-rendering application

A system for medical image processing has been proposed, which allows multimodal dynamic three-dimensional (3-D) visualization interactively and in real time. The system has been conceived to support medical specialists in the diagnosis of moving organs, such as the heart during the cardiac cycle, allowing them to compare information on perfusion/contraction match as a basis for diagnosis of important cardiovascular diseases. The 3-D volume rendering algorithm runs on a SIMD machine because of the great amount of data to be manipulated by always using the same operations. One of the features of the algorithm is the possibility to change, interactively, image processing and visualization parameters at any step, and to perform simple and effective image manipulations. Performance studies have demonstrated a very high global efficiency in practical situations by using typical data-volume dimensions. The system has been tested in the medical environment, by using magnetic resonance (MR) and single-photon emission-computed tomographic (SPECT) images     

Blind image quality assessment with channel attention based deep residual network and extended LargeVis dimensionality reduction

Image Quality Assessment (IQA) is one of the fundamental problems in the fields of image processing, image/video coding and transmission, and so on. In this paper, a Blind Image Quality Assessment (BIQA) approach with channel attention based deep Residual Network (ResNet)and extended LargeVis dimensionality reduction is proposed. Firstly, ResNet50 with channel attention mechanism is used as the backbone network to extract the deep features from the image. In order to reduce the dimensionality of the deep features, LargeVis, which is originally designed for the visualization of large scale high-dimensional data, is extended by using Support Vector Regression (SVR) to perform on a single feature vector data. The extended LargeVis can remove the redundant information of the deep features so as to obtain a low-dimensional and discriminative feature representation. Finally, the quality prediction model is established by using SVR as the fitting method. The low-dimensional feature representation and quality score of the image form the pair-wise data samples to train the fitting model. Experimental results on authentic distortions datasets and synthetic distortions datasets show that our proposed method can achieve superior performance compared with the state-of-the-art methods.     

P2P网络计算中的数据存储研究

Peer-to-Peer（P2P）计算是一种正在迅速发展和不断成熟的技术，在网络计算环境与分布式处理系统中，P2P已经成为一种重要的系统结构：在结构上，它具有很好的可扩展性和灵活性，一般不存在中心管理控制节点。讨论了P2P计算环境中的数据存储与管理，特别是对数据存储的要求以及所面临的挑战进行了分析，并指出了P2P研究与发展中的一些方向。     

Computation of Storage Requirements for Multi-Dimensional Signal Processing Applications

Many integrated circuit systems, particularly in the multimedia and telecom domains, are inherently data dominant. For this class of systems, a large part of the power consumption is due to the data storage and data transfer. Moreover, a significant part of the chip area is occupied by memory. The computation of the memory size is an important step in the system-level exploration, in the early stage of designing an optimized (for area and/or power) memory architecture for this class of systems. This paper presents a novel nonscalar approach for computing exactly the minimum size of the data memory for high-level procedural specifications of multidimensional signal processing applications. In contrast with all the previous works which are estimation methods, this approach can perform exact memory computations even for applications with numerous and complex array references, and also with large numbers of scalars     

大数据处理技术基于云计算的探讨

随着社会经济的不断进步,计算机网络技术得到前所未有的快速发展,其中大数据更是渗透到现代社会的各个领域,给人们的生产生活带来了极大的便利.近年来,高速发展的物联网技术、移动互联网技术等使得网络终端产生了海量的网络数据,数据处理压力十分巨大.基于这种大数据存储和处理需要,网络云计算技术应运而生并且得到了巨大发展.对此,本文在分析云计算的基础上对大数据处理技术进行了重点探讨.