移动IPv6网络家乡代理容错方案性能分析

移动IPv6网络家乡代理的容错机制可保证在家乡代理失效后,节点仍能被寻址。针对现有容错方案普遍存在缺少容错性能定量评价的问题,对其工作机制进行深入研究,并通过仿真对不同方案的容错性能进行定量分析和评价。仿真结果表明：环状备份方案（MADB）在容错时间、服务中断时间以及信令开销等方面的性能均优于其他方案,综合性能提高35%左右。     

异构备份式的虚拟网映射方法研究

在云计算和数据中心环境中,底层单个物理服务器的失效将对上层虚拟网络的服务性能造成很大的影响,现有利用冗余备份的方法能够在一定程度上降低底层物理设备失效带来的影响,但未考虑到物理服务器的同构性所带来的问题,为此,该文提出一种异构备份式的虚拟网映射方法。首先,只对关键的虚拟机进行冗余备份,降低备份资源的开销;然后,确保提供备份虚拟机的物理服务器与原物理服务器的系统类型的异构性,提高虚拟网的弹性能力;最后,以最小化链路资源开销作为虚拟网的映射目标,进一步降低备份资源的开销。实验表明,该方法在保证虚拟网络映射性能的前提下,能够大大提高虚拟网络的弹性能力。     

一种距离扩展目标的Hough变换检测器

针对宽带雷达信号得到的距离扩展目标在有较大运动速度时积累检测困难的问题,该文提出一种基于Hough变换的检测器(HTD)。利用距离扩展目标高分辨距离像信号的相邻回波之间信号具有较大相关系数这一特点,在互相关序号与相关时间2维平面使用Hough变换进行积累来检测目标。理论推导表明Hough变换检测器不需要目标散射中心分布的先验知识和目标的运动信息,具有恒虚警(CFAR)性能。计算机仿真实验表明Hough变换检测器具有比不依赖散射中心分布信息的广义似然比(NSDD-GLRT)检测器更好的检测能力。     

基于联合备份的服务功能链可靠性保障的部署方法

在网络功能虚拟化(NFV)环境中,针对服务功能链(SFC)部署时的可靠性问题,该文提出对备份虚拟网络功能选择、备份实例放置和服务功能链部署的联合优化方法。首先,定义一个单位开销可靠性提高值的虚拟网络功能衡量标准,改进备份虚拟网络功能选择方法;其次,采用联合备份的方式调整相邻备份实例之间的放置策略,以降低带宽资源开销;最后,将整个服务功能链可靠性保障的部署问题构建成整数线性规划模型,并提出一种基于最短路径的启发式算法,克服整数线性规划求解的复杂性。仿真结果表明,该方法在优先满足网络服务可靠性需求的同时,优化资源配置,提高了请求接受率。     

一种基于结构化备份子图的路由系统失效恢复方法

鉴于当前路由系统失效恢复未能有效解决备份拓扑存储代价、冗余恢复能力以及自治域利益保护等问题，该文在致瘫攻击易造成网络级联失效的背景下提出一种基于结构化备份子图的失效恢复方法3R(Robust Route Recovery)。首先，设计拓扑关键点和重要邻接点生成算法，满足了备份子图基数小、增长率低的实际存储需求，且具备同一子图内多点冗余恢复特性；其次，为充分考量自治域利益诉求，设计基于流量权重的相邻链路排序算法，折中路由失效恢复与私有路由策略的双重需求；在此基础上，利用多次迭代生成各冗余恢复集合的备份子图。实验结果证实了方法的有效性。     

WLAN中分布式入侵检测系统研究

文中通过分析WLAN的特点及存在的安全问题,提出可以用于基础结构模式WLAN的分布式入侵检测系统,该系统主要由中心控制台和监测代理组成,中心控制台配置监测代理和浏览检测结果,并进行关联分析。监测代理监听无线数据包、利用检测引擎进行检测、记录警告信息,并将警告信息发送至中心控制台。实验表明,该模型对WLAN的安全保障具有一定的实用价值。     

一维均值反差滤波的小目标快速检测算法

复杂背景下低信噪比弱小目标的自动检测是当今目标自动探测研究尚未解决的一个难题.将一维均值反差作为一种不相似性度量应用于小目标检测的前期滤波处理中,有效地增强小目标信息、抑制了复杂的背景和噪声,并结合背景预测原理,实现了对小目标的快速检测.仿真实验表明:该滤波算法大幅提高了小目标图像的信噪比,保证了利用背景预测原理检测小目标的准确性;与基于二维均值反差的滤波方法相比,该方法对小目标形状的适应性更强,速度更快.     

联合瞬时参数分析用于探地雷达目标增强

通过对探地雷达（GPR）信号的瞬时幅度、瞬时相位和瞬时频率特征进行分析,发现GPR信号的瞬时幅度信息反映了分界面的埋深和各层介质对发射脉冲信号的吸收能力;其瞬时频率信息对晚时信号具有增强作用,并反映了相邻介质介电属性之间的强弱关系。在此基础上,提出了利用瞬时幅度与归一化瞬时频率相乘积的联合瞬时参数方法对GPR信号进行预处理,以提高相邻介质分界面的定位精度。仿真实验和野外实测数据实验结果表明：该方法具有较高的分辨率,能精确地定位相邻介质分界面对应的时刻,其目标增强性能优于Hilbert谱分析方法。     

基于备份的可重构服务承载网可靠性映射方法

可重构柔性网络链路失效将严重影响其上承载的可重构服务承载网(RSCN)的可靠性。文章基于路径备份策略着重解决时延敏感类型RSCN的可靠性问题,并提出分阶段处理方式进一步优化备份资源消耗。在拓扑预处理阶段,根据RSCN是否支持路径分裂分别提出分裂的最小备份拓扑生成(S-MBT-Gen)算法和最小备份生成树(MBST-Gen)算法,减小备份拓扑带宽约束总量;在拓扑映射阶段,提出主备拓扑协同映射(RNM-PBT)算法,协调利用底层网络资源。仿真结果表明,本文提出的算法降低了RSCN平均资源消耗,且具有较高的请求接受率和较低的平均执行时间。     

三平行线阵中基于改进传播算子的二维DOA估计方法

提出了一种三平行线阵中基于改进传播算子的二维波达方向估计方法。该方法针对三平行阵列的结构特点,利用三平行线阵中两个相互垂直的双平行线阵,分别构造一个扩展传播矩阵,求得旋转矩阵。然后根据三平行线阵的特性,对分别得到的旋转矩阵进行配对。最后利用配对后的旋转矩阵,联合估计信号的方位角和俯仰角。该方法当俯仰角在70°~ 90°之间时不会出现角度估计失效问题,充分利用了所有阵元信息,提高了角度估计性能,而且具有更低的计算复杂度。仿真验证了提出方法的有效性。      

Spatial validation of the collection 4 MODIS LAI product in eastern Amazonia

This paper reports on the validation of the Collection 4 MODIS leaf area index (LAI) product over the Tapajo/spl acute/s region, eastern Amazonia. The validation site is enclosed in tile h12v09 of the MODIS LAI product. The methodology to assess MODIS LAI accuracy included two main steps: (1) a multiple regression analysis for the generation of LAI surfaces, based on the relationships between field data and remote sensing information from the Enhanced Thematic Mapper Plus sensor, and between field data and topographic information from a digital elevation model; (2) the direct comparison of these LAI surfaces with the MODIS LAI surfaces. The analysis indicated that MODIS LAI is significantly overestimated for the Tapajo/spl acute/s region by a factor of 1.18. No relationships between MODIS LAI and the validation surfaces were found. These results are indicative of a predominance of LAI retrievals by the backup algorithm, which is overcompensating LAI values at the saturation domain. The overgeneralization of the land cover layer (MOD12Q1) can be a source of uncertainties for the lookup table parameterization. Further validation efforts must be carried out over Amazonia for a quantitative quality assessment of the MODIS LAI surfaces in order to improve its accuracy.     

基于网络的敌我识别方法初探

为提高复杂战场环境下的敌我识别效能,提出了一种基于网络的敌我识别信息分发共享新方法.分析了具有异构自治、动态临时的敌我识别网络特性;并对影响信息分发共享过程抗毁性的链路备份能力和节点备份能力进行了研究,从完整性、正确性和时效性3方面说明了方法的有效性.结果表明,该方法对识别效能的提高具有积极作用.     

一种以备份信息为基础的译码新方法

提出了一种新的译码方法———"备份信息"译码方法,该方法在编码的"备份信息"全集中适当选取子集并加以组合,最终形成判决量进行译码。此方法在保证误码率基本不变的情况下大大减少了译码复杂度。计算机仿真表明这种译码方法在短码译码上有较好的性能。     

A Dynamic Backup Path Management Method for TDMA Based WiMAX Client WMNs

This paper proposes a backup path management method for time division multiple access (TDMA) based client wireless mesh networks (WMNs). In a TDMA based client WMN, as links/nodes fail or as nodes perform handover and as flows enter and leave the network, the paths between various nodes change as well as the bandwidth available along these paths. In these networks, to support the quality of service requirements of flows, backup paths with the required bandwidth need to be established dynamically. Some methods are proposed in the literature to establish backup paths which handle link/node failures and node handover in ad hoc networks, but none of these methods can provide backup paths with the required bandwidth dynamically. To address that issue, the present paper proposes a backup path management method which is adaptive to both topological changes and traffic changes in a network. Each node along the current path between a source and a destination finds backup paths with the required bandwidth in order to handle failure of the link to its downstream node and its own failure or handover. Nodes use two-hop neighborhood information and slots status information of two-hop neighbors to establish backup paths. We prove that the number of backup paths available when a node N searches for backup paths to handle its own failure are more than the number of backup paths available when some other node searches for the backup paths for the failure of node N. Performance of the proposed method is compared with the performance of a naive path management (NPM) method in which always the source establishes backup paths whenever a link/node fails or a node performs handover, and also with the performance of a backup path management method proposed in the literature. The proposed method significantly outperforms the NPM method and the method selected from the literature. For example, when the speed of the mobile nodes is 50 m/s, the packet delivery ratio with the proposed method is 63 % more than the NPM method and 35 % more than the method selected from the literature.     

基于医院网络的交换机配置文件批量备份技术

在数据信息时代中,智能化的数据信息网络成了医院基础信息业务的重要组成部分,随着医院网络化信息建设业务的不断深化,基础数据信息的稳定性变得至关重要。医院的网络信息系统主要由交换机、汇聚交换机以及数台接入层组成,在一线的日常管理中,需要强化对基础数据信息的维护与备份,最大限度地保障医院基础业务的开展。文中首先综述了医院交换机硬件配置备份的场景,其次阐释了交换机数据处理中的SNMPv2c协议与TFTP协议,并从网络拓扑结构设计以及基于WinSock控件的配置备份的代码实现中,总结了医院网络交换机的定期自动备份方法,还给出了控制点发生错误时的应对策略,希望能为医院计算机网络的交换机优化配置与信息化管理提供参考和借鉴。     

Enhancing Security in Mobile IPv6

In the Mobile IPv6 (MIPv6) protocol, a mobile node (MN) is a mobile device with a permanent home address (HoA) on its home link. The MN will acquire a care‐of address (CoA) when it roams into a foreign link. It then sends a binding update (BU) message to the home agent (HA) and the correspondent node (CN) to inform them of its current CoA so that future data packets destined for its HoA will be forwarded to the CoA. The BU message, however, is vulnerable to different types of security attacks, such as the man‐in‐the‐middle attack, the session hijacking attack, and the denial‐of‐service attack. The current security protocols in MIPv6 are not able to effectively protect the BU message against these attacks. The private‐key‐based BU (PKBU) protocol is proposed in this research to overcome the shortcomings of some existing MIPv6 protocols. PKBU incorporates a method to assert the address ownership of the MN, thus allowing the CN to validate that the MN is not a malicious node. The results obtained show that it addresses the security requirements while being able to check the address ownership of the MN. PKBU also incorporates a method to verify the reachability of the MN.     

核心链路感知的可生存虚拟网络链路保护方法

针对现有可生存虚拟网络链路保护方法无差别对待所有虚拟链路、备份资源消耗多且故障后网络恢复时延长的问题,该文提出一种核心链路感知的可生存虚拟网络链路保护(CLA-SVNLP)方法。首先,综合考虑虚拟链路动态和静态两方面因素构建虚拟链路核心度度量模型,依据虚拟网络生存性需求,对核心度较高的虚拟链路进行备份保护;其次,将p圈引入可生存虚拟网络链路保护,依据虚拟网络特点构建p圈,为核心虚拟链路提供1:N保护,即每条核心虚拟链路平均消耗1/N条的备份链路带宽资源以减少备份链路资源消耗,并将单物理链路保护问题转化为多个p圈内的单虚拟链路保护问题;最后网络编码技术与p圈结合,将备份链路对核心虚拟链路提供的1:N保护转化为1+N保护,避免了故障后定位、检测及数据重传。仿真结果表明,该方法提高了备份资源利用率且缩短了故障后的网络恢复时延。     

Adversarial image detection based on the maximum channel of saliency maps

Studies have shown that deep neural networks (DNNs) are vulnerable to adversarial examples (AEs) that induce incorrect behaviors. To defend these AEs, various detection techniques have been developed. However, most of them only appear to be effective against specific AEs and cannot generalize well to different AEs. We propose a new detection method against AEs based on the maximum channel of saliency maps (MCSM). The proposed method can alter the structure of adversarial perturbations and preserve the statistical properties of images at the same time. We conduct a complete evaluation on AEs generated by 6 prominent adversarial attacks on the ImageNet large scale visual recognition challenge (ILSVRC) 2012 validation sets. The experimental results show that our method performs well on detecting various AEs.     

可在线学习的视频目标检测系统

本文针对视频图像中的目标检测问题,基于在线学习方法,研究具有自主学习能力的目标检测系统。该系统由目标检测模块及检测结果验证模块组成,目标检测模块由具有在线学习能力的分类器构成,检测结果验证模块通过粒子滤波对系统检测到的目标估计其后验概率分布,从而验证该检测结果是否为真实目标,并从中获得在线学习样本,实现检测系统的无需人工干预的自适应学习。为了减少验证错误对系统在线学习的影响,提出基于多信息融合的粒子滤波验证方法提高系统的鲁棒性。本文对PETS2006视频序列以及公交车内视频序列进行了目标检测实验,证明了其较强的自适应能力和较好的检测效果。      

Low-cost binding update scheme based on pointer forwarding in HMIPv6 network

By introducing a mobility anchor point(MAP),hierarchical mobile IPv6(HMIPv6)reduces the binding update signaling cost associated with mobile IPv6,but there still exist deficiencies.For instance,a mobile node(MN)needs to orderly accomplish two binding updates with the MAP and home agent(HA)when the MN performs inter-MAP mobility.This results in a high signaling cost,thus affecting network performance.To reduce the inter-MAP binding update cost of idle MN in HMIPv6,an optimization scheme based on pointer forwarding with a threshold is proposed.The scheme can reduces the binding update cost of idle MN by using the binding update between MAP to replace several home binding updates.The signaling cost difference is derived by analyzing the cost of the basic scheme and the optimization scheme between two successive sessions.Simulation results show that,the optimization scheme can reduce the binding update signaling cost and improve the network performance as long as a suitable threshold is chosen.The discussions on the sensitivity of tele-parameters are also given.