基于Linux/QT的火控系统实时故障诊断方法

针对某数字化火控系统实时故障检测与诊断问题,提出了一种基于对象式框架知识表示的火控系统实时故障诊断方法。该方法采用专家系统与实时进程通信,获取故障征兆,进行实时推理定位故障源。给出了诊断系统的一般架构,介绍故障诊断系统知识表示的形式化描述,实时推理、规则匹配算法和UDP进程通信协议。实际应用证明了这种方法的有效性和可扩展性。     

基于非负张量表示的扩展LDA标注模型

由于“语义鸿沟”的存在,自动图像标注是一项极具挑战性的工作。考虑到图像低层视觉特征与高层语义概念的差异,分别从图像表示与语义建模两个方面来实现自动图像标注。在图像表示方面,提出了一种正则化约束下的非负张量表示方法,用以提取符合人眼视觉直观理解的图像高阶结构特征。在语义建模方面,提出了一种三层贝叶斯模型——扩展隐Dirichlet分配。该模型利用隐变量来实现图像与标注词的关联,并通过一种基于变分推理的期望最大值方法来估计其参数。实验结果表明,ELDA模型在大规模数据库NUS-WIDE上的标注结果相较于现有方法有了显著的提高。     

新一代指挥信息系统中的语义网应用能力分析

随着现代战争形式的演变,提升指挥信息系统能力的关键在于信息处理.语义网是一种先进的信息表示方法,能够有效地提升机器的信息理解和处理能力.简要介绍了语义网技术的组成和本质能力,并结合美军的研究实例,分析了其在新一代指挥信息系统中的应用能力,指出了发展机遇和挑战,并对未来研究提出了建议.     

基于知识的模型表示研究

基于知识的模型表示是信息管理系统,特别是决策支持系统(DSS)的基础。根据模型管理和模型建立动态性原则的要求,提出了模型知识化的框架表示法,并给出了应用实例。     

‘God willing,I will be back’: Gauging the Truth and Reconciliation Commission's capacity to deter economic crimes in Liberia

Recently, in July 2012, the high-profile and bitterly fought nine-month race for the post of Chair of the African Union (AU) Commission, between Dr Jean Ping of Gabon, and his main challenger, Dr Nkosazana Dlamini-Zuma, culminated in the latter's victory. Her victory came after the deadlock at the Eighteenth AU Summit in January 2012. Borne out of the considered need for a quick fix through reliance on a vote of expediency, the election of Dr Dlamini-Zuma represented a political resolution to the crisis that arose due to the earlier ongoing electoral deadlock. Far from being a competitive election by design, the 2012 AU Commission election by default became transformed into an intensely fought campaign that put the AU in the limelight. This article briefly introduces the electoral process, explains in detail the voting behaviour of AU member states, and offers five specific reasons for the victory of Dr Dlamini-Zuma. On the surface the election looks very competitive, but the article explains why this is not the case. To create greater competition for these posts, the AU needs to overhaul the nomination process and the voting procedure. In this regard, the article proffers detailed analysis and proposes a radical revision of the existing criteria for the nomination. The article also proposes specific recommendations for the amendment of the rules of procedure of the AU Assembly to allow for a qualified majority as a deadlock breaker in the fifth round. It also assesses whether the integrity of the AU Commission election was damaged during the campaigning and voting process. In this regard, it recommends the development of a code of conduct for future elections at the AU.     

THE AU MISSION IN BURUNDI

Abstract

Even though the Movimento Popular de Libertação de Angola (Popular Movement for the Liberation of Angola, MPLA) managed to take firm control of the state as the war ended in 2002, this article contends that its hegemony is not absolute. At the fringes of the state administration, emerging regional elites have established associations that seek to provide a remedy against the centralised system, which essentially deprives regions of political influence and sufficient economic resources. As in many other countries on the African continent, political identities that were legally enforced and institutionally reproduced in colonial times have hardly been transcended in the post-colonial period. Even if the associations' influence today does not yet reach further than bringing some insecurity to an otherwise secure polity dominated by the MPLA, the associations could play a role in reinforcing strong regional solidarity and give political expression to feelings of exclusion. Such sentiments could in turn lead to an increase of exclusivity ideas about citizenship, or even xenophobic violence. Much will depend on the strategies of the MPLA, and to what extent it will accommodate these emergent forces in the post-war era, not the least through the current sub-national state reforms.     

应用局部约束二维稀疏表示识别SAR图像目标

在分析典型稀疏表示分类方法和局限性的基础上,提出了一种基于局部约束的二维稀疏表示方法,以有效解决SAR图像目标识别问题。该方法对SAR图像进行图像预处理,在兼顾图像相邻列(行)对应稀疏表示系数邻近性和样本间局部性的基础上,构建了局部约束目标函数,并通过解闭式解,实现稀疏表示系数的更新求解。利用美国实测MSTAR数据对算法进行了仿真验证,实验结果表明所提出的方法可实现SAR图像目标的有效识别,并对训练样本数目具有一定的鲁棒性。     

基于四阶累积量阵列扩展的稀疏分解方法

提出了一种基于四阶累积量稀疏表示的估计方法,解决信号数多于阵元数时的DOA估计问题。该方法首先构造了包含所有DOA信息的最小冗余矢量,利用扩展阵列的最小冗余导向矢量构造完备字典减小完备字典的复杂度;然后利用L1范数作为稀疏约束条件建立稀疏模型进行DOA估计。理论分析和仿真实验,验证了该方法能够估计出的信号个数大于阵元数目,可直接应用于相干信号,比MUSIC-like算法具有更好的性能。     

多任务时间规划冲突检测方法

摘要：针对多任务的时间规划问题,基于Simple Temporal Networks(STN)时态表示模型和时态推理方法,提出了一种时间冲突检测的智能方法,并对该方法给予了数学证明。实验表明,该方法能够检测出基于STN表示的作战方案的时间冲突,并对任务时间规划提供参考范围,可有效辅助指挥员快速制定作战方案的时间规划。     

转动算符表示矩阵的计算

利用量子体系态空间基矢的完备性,通过表象变换,直接计算出转动算符∧D(α,β,γ)在{∧J~2,∧J_z}表象中的矩阵表示.该方法简单而有效.