NS2扩展突发性业务流：设计与实现

结合Markov调制泊松过程（MMPP）和混合泊松流,为突发性分组业务设计了Gamma调制泊松过程（GMPP）业务流模型。面向开源仿真器NS2的功能扩展,给出了纯泊松流、MMPP流和GMPP流的NS2扩展设计和实现方案,并经仿真实验进行了验证。分析表明GMPP更适用于实际业务流。     

蜂窝移动网中基于MMPP业务输入模型的系统性能分析

文章针对现有的以Poisson过程作为输入业务流模型分析系统性能的不足,提出采用马尔可夫调制泊松过程(MMPP)作为输入业务流来分析系统的呼叫阻塞率和掉话率.数值和仿真结果表明,MMPP能够较好的对蜂窝移动网中的业务输入流近似建摸,它考虑了业务流本身的相关特性和突发特性,可以更加准确的用于分析系统性能.     

SAR图像的Gamma混合分布建模方法

随机信号的混合概率模型比单一概率模型具有更多的灵活性,更适合复杂的分布建模。当前主要的混合概率模型有高斯混合模型、α分布混合模型和Gamma混合模型等。但高斯混合模型更适合随机变量对称分布的分布建模,而α混合模型参数多、算法复杂。SAR图像的像素值为非负值,且多为斜峰分布,更适合用Gamma混合模型建模。仿真分析及数据测试都表明,本文提出的gamma混合分布建模方法对SAR图像的像素统计分布具有更高的运算效率。      

基于混合Gamma建模与MRF的SAR图像分割方法

文中基于结合利用SAR图像统计信息和像素的空间约束信息的思路,提出了基于混合Gamma建模和MRF的分割方法,算法中利用混合Gamma模型对SAR图像进行统计建模,利用MRF模型对像素空间相关性进行建模。文中验证了混合Gamma模型参数估计的准确性以及自适应估计分割类别数目的有效性。最后用模拟SAR数据和实测SAR数据验证了文中所提出的分割算法的有效性。     

两跳中继系统在混合Gamma衰落信道下的性能分析

基于混合Gamma信道模型研究了两跳中继系统在联合多径阴影衰落环境下的性能.推导了两跳中继系统端到端信噪比的累积分布函数和矩生函数的精确闭式解,并基于这些闭式解分析了两跳中继系统的误码率、中断率和分集增益等性能.与广义K衰落模型下的性能进行了比较,运用混合Gamma衰落模型可获得简单精确的结果.最后数值和仿真分析验证了理论研究的正确性.     

基于MATLAB的RMD算法模型设计

现有的业务流模型难以描述高速业务流的高突发性,在线测量网络中的高速业务流,再根据实测的时间序列数据,用RMD方法建立并分析高速业务流的自相似模型。RMD模型自相关函数的近似程度较好,可以较好的模拟业务量。     

光突发交换中的突发业务流模型及其应用

提出了一种光突发交换中的突发业务流模型,采用该模型对光突发交换中的LAUC-VF输出调度算法在不同的突发业务强度和突发长度下的性能进行了模拟仿真,分析比较了该算法在此突发业务流和普通业务流模型下的性能,仿真结果表明,该突发业务流模型具有一定的合理性。     

网络业务流模型综述

通信网络需要提供满足用户需求的QoS(quality of service)，这就需要业务流模型能准确地捕获实际业务流的统计特征。业务流模型在网络的流量控制及性能分析中都起至关重要的作用。随着网络业务类型不断增加，实际网络中业务流特性越来越复杂，表现为突发性、长相关性或自相似性等。这使得业务流建模方法的研究得到不断发展。文中就当前网络的业务流特性，分析了现有的业务流模型，并指出了这些建模方法的特点。     

基于MATLAB的TES建模软件设计

现有的业务流模型难以描述高速业务流的高突发性。在线测量网络中的高速业务流，再根据实测的时间序列数据，用TES(Transform-Expand-Sample)方法建立并分析高速业务流的TES模型，其中关键环节是TES模型的建立.在TES模型构架的基础上，给出了基于MATLAB平台的有关的软件实现方案.     

一种基于混合Gamma分布的自动目标识别混合EM算法

提出一种基于高分辨距离像的混合EM算法用于自动目标识别,采用混合Gamma分布对目标主要散射点的特征进行建模。混合Gamma分布的参数分为两类:1)可以通过EM算法进行更新Gamma分布的混合系数和尺度参数;2)不存在解析更新表达式的Gamma分布的形状参数。提出的算法可为两部分:1)原始EM算法的加速算法;2)利用共轭梯度法估计Gamma分布的形状参数,并对更新表达式进行推导。数值仿真结果验证了提出方法的有效性。     

一种基于扩展自相似过程的VBR视频业务新模型

本文在扩展自相似过程的基础上 ,提出了一种简单的视频业务新模型。该模型的边缘分布近似Gamma分布 ,且其具有长或短的相关性。在许多时间尺度上 ,这些特性可以很好地匹配众多实际变比特率视频业务的一、二阶统计特征。仿真实验中 ,通过利用该模型模拟实际视频信号 ,我们证明它也能很好地反映实际视频业务的排队性能     

An Analytical Formulation for Grade of Service Determination in Telephone Networks

The paper provides a simple method for point-to-point blocking estimation in telephone networks. A one-moment model is developed which incorporates the definition of fictitious offered traffic that enables one to take into account the deviation of smooth and peaked traffics from the Poisson. Numerical results illustrate the accuracy of the method.     

多媒体通信中不同优先级的业务复用模型分析

本文建立了基于流体流方法的复用模型，分析了具有不同优先级的多业务复用情况，计算了平均时延、平均队长及分组丢失率等服务质量参数，比较了有无优先机制时上述服务质量参数的异同。     

Analysis of a model of two overflow traffic components with different serving intensities

This paper seeks to provide the analysis of a model of two overflow traffic components that originated from several Poisson traffics served by a common primary trunk group. Overflow parts of two of them are served in a secondary trunk group, but with changed and different serving intensities. The general solution of the system was obtained by generating function techniques, and call losses, time losses and traffic losses were defined based on state probability and solved numerically. In order to illustrate specific characteristics of the analyzed model, loss values were compared with the losses in a corresponding system with averaged serving intensity for which there is a much simpler explicit solution.     

移动因特网不同流量下切换损失率研究

根据移动因特网中基本的硬切换和改进的半软切换算法时间流程仿真研究了 2种切换在泊松和自相似流量下的切换损失率。仿真数据表明 ,在网络负载均值相同的情况下 ,对于硬切换和半软切换 ,自相似流量的切换损失率要明显低于传统泊松流量 ,并通过分析泊松和自相似流量的概率密度函数尝试给出产生这一区别的原因。     

Solving some overflow traffic models with changed serving intensities

This paper investigates specific systems with overflow traffic. A primary group with two Poisson traffics is considered whereupon the rejected calls from one traffic are directed to the alternative group with changed serving intensities. The generating function technique is used for analytical solving the model with secondary and ternary groups and the model that separately treated the channels in alternative groups. The obtained analytical solutions essentially reduce the constraints concerning the equation system size, convergence, and calculation time, which arise when numerically solving the steady-state system equations. For the case with single channels in the ternary group, explicit solutions for traffic parameters are obtained. Also, comparison with the model that has a unique serving intensity of overflow traffic is made.     

基于PGBN模型的SAR图像目标识别方法

特征提取是合成孔径雷达图像目标识别的关键步骤,也是难点之一。该文提出一种基于PGBN(Poisson Gamma Belief Network)模型的SAR图像目标识别方法。PGBN模型作为一种深层贝叶斯生成网络,利用伽马分布具有的高度非线性,从复杂的SAR图像数据中获得了更具结构化的多层特征表示,这种多层特征表示有效提高了SAR图像目标识别性能。为了获得更高的训练效率和识别率,该文进一步采用朴素贝叶斯准则提出了一种对PGBN模型进行分类的方法。实验采用MSTAR的3类目标数据进行了验证,结果表明通过该方法提取的特征有更好的结构信息,对SAR图像目标识别具有较好的性能。     

Impulsive noise measurements and characterization in a UHF digitalTV channel

This paper presents the results of a study covering measurement and characterization of the wide-band impulsive noise present in a digital TV radio channel. Measurements were conducted at a frequency of 762 MHz in different outdoor and indoor environments using vertical and horizontal polarization. The measurement system was built on commercial equipment only. The calibration process, which is an important stage of this kind of measurements, is described. To analyze the measurements the impulsive noise has been modeled as a pulse train where the pulse amplitude, pulse duration and elapsed time between pulses are considered random variables. It has been found that the pulse duration and elapsed time between pulses is not dependent on the antenna polarization while the pulse amplitude is, especially in the case of the noise generated by a fluorescent lamp. It has also been found that the pulse duration of the noise measured in the outdoor environments presents some clustering features and is correlated with the pulse amplitudes. This correlation may be caused by a RF noise bandwidth that is larger than the bandwidth of the measurement system. The noise in busy streets presents larger pulse durations, larger amplitude, and shorter elapsed time between pulses that the noise measured in a pedestrian area. Several statistical tests have been done to find the distribution function that best fits these random variables. Power Rayleigh, lognormal, exponential, Poisson, and Gamma distributions have been tested. According to the assessment carried out, none of the distribution functions is adequate to model the pulse amplitudes or the elapsed time between pulses, while the pulse duration seems to be Gamma distributed     

MPEG标准VBR视频业务新模型及其性能研究

迄今为止,MPEG-1标准下的可变比特率(VBR)视频业务建模问题还没有得到很好的解决.鉴于此,本文在分析这类业务统计特性的基础上,对这类业务提出了一种新的帧层模型.该模型可以很好地体现这类业务的一、二阶统计特性.通过使用该模型模拟实际视频信号,也显示了该模型在体现实际业务排队性能上的有效性.