基于ZIGBEE网络的ZiCL改进算法

针对ZigBee网络中使用的ZiCL算法可能造成节点过度耗能和网络能耗不均的问题,提出一种改进算法。该算法根据网络运行过程中簇成员能量状态来动态调整路由选择策略。仿真结果表明,改进算法提高了网络传输的可靠性,有效节约了能量。     

正则蕴涵算子族G-λ-R0及其三Ⅰ支持算法

首先给出了一个新的蕴涵算子族:G-λ-R0(λ∈[0,1])(它包括Gsdel(简称RG)算子与R0算子).然后重点讨论了G-λ-R0(λ∈[0,1])族算子的伴随算子及其正则性.结果表明,在该算子族中,每一个算子都具有伴随算子且具有正则性.从而说明了此算子是较理想的蕴涵算子.最后讨论了基于此蕴涵算子族的三Ⅰ支持算法.     

Main and interactive effects of shiftwork, age and work stress on health in an Italian sample of healthcare workers

Among healthcare workers, shiftwork (mostly if nightwork is also included), ageing and work-related stress may be factors leading to impaired health. Such risk factors may also operate in interaction, resulting in an even increased harm for health. The present study aims at evaluating these relationships in a sample of 1842 hospital workers in Northern Italy. Subjects were mainly women, 33.1% were aged > or = 45 yr, and they were almost evenly distributed between dayworkers and rotating shiftworkers (nights included). Shiftwork was associated with poor sleep, while it was protective against gastrointestinal disorders, poor work ability and job dissatisfaction. Work stress was the risk factor with the highest relevance for poor health. Ageing was associated with lower physical health. Few significant interactions were observed. Shiftwork with nights and high work stress significantly interacted in increasing the risk for poor sleep. The "healthy worker effect" may have played a strong role in study findings.     

Web系统性能测试研究

Web系统的性能测试是Web应用开发过程中的一个关键环节,然而Web站点的复杂性及其用户行为的不可预见性使得对其性能测试极其困难,而且测试环境和用户使用环境的差异性使得Web系统性能测试数据不精确或不可靠。分析了Web系统结构和Web系统的性能特点,深入讨论了Web系统性能测试的内容和性能测试的原理,并比较了目前常用的性能测试工具,通过一个测试项目介绍了自动化测试的实施过程,总结出一套切实可行的Web系统性能测试流程。     

表面滚压强化工艺的数值仿真

针对圆棒工件表面进行多圈连续进给滚压强化工艺,提出了Mare有限元的三维数值仿真模型.通过边界条件施加、滚动的实现、接触设置等有限元建模的建屯和修正,取得了与实际十分吻合的滚压成形的仿真结果;在分析残余应力形成过程的基础上,提出了改进的滚压过程模型;比较了不同进给量、网格划分及边界条件对残余应力分布的影响规律.结果表明,与以往的单圈或者不足一圈的滚压、滑动滚压,或者轧辊不转动的静止压入等简化条件下的仿真模型相比,仿真模型更接近实际的滚压过程,因而也获得了比较理想的变形和残余应力的仿真效果.     

On the investigation of the neurophysiological correlates of knowledge worker mental fatigue using the EEG signal

Technological trends and advances in automation have underscored the importance of task performance of certain jobs requiring mental functions such as information processing and decision analyses. Most experts agree that such work environments produce increased mental activities, with profound implications for mental fatigue and stress. Consequently, productivity measurement and improvement for white collar or ‘knowledge worker’ occupations remains a major challenge and concern. This investigation defines an experimental approach that examines the neurophysiological correlates of white collar worker mental fatigue using the EEG signal. A 6 h laboratory experiment was conducted to simulate work output. The methods of assessing fatigue employed were mental tests and physiological measurements. The experiment involved reading of standardized texts, finding solutions to arithmetic-logical problems and a combination of both task types. Two primary performance measures were obtained, work output and brain waves. Fast Fourier transform and correlation analyses are used to quantify the relationship between certain brain waves and mental fatigue. This research is a major step towards the development of a model that explores the relationship between mental fatigue and factors associated with output performance, optimal recuperation periods and related variables. Such a model would be useful in human reliability prediction based on task parameters and worker profiles.     

基于残差神经网络的无人机射频信号识别

针对民用小型无人机对低空领域安全造成威胁及无人机视觉上识别难的问题,提出了一种无人机射频信号识别的方法,使用深度学习技术学习无人机与控制器之间的射频信号特征来识别无人机。首先将射频信号数据集中的无人机信号进行预处理操作,接着使用残差神经网络进行模型训练,最后使用训练好的网络模型对无人机信号进行识别验证。实验结果表明,该方法识别无人机是否存在的准确率达到99.8%,识别无人机型号的准确率达到91.1%,识别无人机运行模式的准确率达到70.3%,且该方法具备较强的鲁棒性和环境抗干扰能力,性能明显优于基准方法。     

基于提取双选紧密特征的RGB-D显著性检测网络

针对现有算法对不同来源特征之间的交互选择关注度欠缺以及对跨模态特征提取不充分的问题,提出了一种基于提取双选紧密特征的RGB-D显著性检测网络。首先,为了筛选出能够同时增强RGB图像显著区域和深度图像显著区域的特征,引入双向选择模块(bi-directional selection module, BSM);为了解决跨模态特征提取不充分,导致算法计算冗余且精度低的问题,引入紧密提取模块(dense extraction module, DEM);最后,通过特征聚合模块(feature aggregation module, FAM)对密集特征进行级联融合,并将循环残差优化模块(recurrent residual refinement aggregation module, RAM)配合深度监督实现粗显著图的持续优化,最终得到精确的显著图。在4个广泛使用的数据集上进行的综合实验表明,本文提出的算法在4个关键指标方面优于7种现有方法。     

基于空间矢量复合判断指标的变电站动力电缆漏电检测算法

为提高变电站长段动力电缆漏电监测水平,提出了一种基于空间矢量复合判据的变电站长段动力电缆漏电检测算法。首先对现有方法中无法检测三相漏电故障的问题进行了分析。然后,引入空间矢量的概念,将剩余电流数据转换为空间矢量圆,提出了剩余电流差流、A相漏电流以及空间矢量圆半径变化率3个漏电状态判断指标。并建立了针对电缆不平衡漏电及三相漏电故障的类型判断机制,实现了对长段动力电缆漏电状态及类型的准确检测。最后,通过仿真及江苏电网某500 kV变电站实际数据对所提方法的有效性与优越性进行验证。结果表明,该方法能够准确实现对不平衡漏电故障、故障演变过程以及三相漏电故障的判断,丰富了电缆漏电故障的诊断信息,有效提高了对变电站漏电问题的分析处理效率。     

基于人脸关键特征提取的表情识别

自然场景下人脸表情由于受遮挡、光照等因素影响,以及表情局部变化细微,导致现有人脸表情识别方法准确率较低。提出一种人脸表情识别的新方法,以ResNet18为主干网络,利用残差连接模块加深网络结构,以提取更多深层次的表情特征。通过引入裁剪掩码模块,在训练集图像上的某个区域进行掩码,向训练模型中增加遮挡等非线性因素,提升模型在遮挡情形下的鲁棒性。分别从特征图的通道和空间两个维度提取表情的关键特征,并分配更多的权重给表情变化明显的特征图,同时抑制非表情特征。在特征图输出前加入Dropout正则化策略,通过在训练中随机失活部分神经元,达到集成多个网络模型的训练效果,提升模型泛化能力。实验结果表明,与L2-SVMs、IcRL、DLP-CNN等方法相比,该方法有效提高了表情识别准确率,在2个公开表情数据集Fer2013和RAF-DB上的识别准确率分别为74.366%和86.115%。