基于DEMATEL和VIKOR的服务外包风险评估

服务外包风险评估是企业供应商管理的重要依据。传统风险评估方法FMEA没有考虑风险模式间的相互影响关系,模糊DEMATEL方法用来分析风险模式的净影响度,进而修正其初始评估的严重度。针对传统FMEA方法将风险因子简单相乘的缺陷,采用模糊VIKOR方法对风险模式的严重度、发生频率和难检度进行综合评估。传统模糊VIKOR计算过程中需要进行模糊数的反模糊化处理,造成信息的损失,相对偏好关系分析引入模糊VIKOR计算解决这一问题。以某公司服务外包风险评估为例,验证了所提方法的有效性。     

基于ICA和EMD的生理信号提取

提出一种新的获取人体生理参数的方法,用摄像头采集人脸的彩色视频,对人脸区域进行颜色通道分离和独立成分分析（ICA）,获取有用信号。使用经验模态分解（EMD）的方法,把信号分解成可以反映出生命信息的固有模态函数（IMF）,再根据所设计的提取准则,分别提取出较为准确的心跳和呼吸信号。用Bland-Altman法进行对比实验分析,结果表明,此方法具有一定的准确性和实用性。     

Ca-Si线与FeS合金化模式对低碳结构钢硫化物形态的影响

通过10 kg感应炉研究了Ca-Si线与FeS加入次序对低碳结构钢(/%:0.16C,0.65Mn,0.25Si,0.034S,0.025P)硫化物形貌和切削性能的影响。以钢中非金属夹杂物变形处理为基础,采用Image Pro Plus软件对钢中夹杂物指标进行定量分析,得出采用先加入Ca-Si线后加FeS的合金化模式时,钢中夹杂物平均长度为12.3μm,单位面积上条状夹杂物与粒状夹杂物之比为64:41,而采用先加FeS后加Ca-Si线的合金化模式,钢中夹杂物平均长度为9.9μm,单位面积条状夹杂物与粒状夹杂物数量之比为52:53,且钢样的切削性能较先加Ca-Si线模式好,因此先加FeS后加Ca-Si线的合金化工艺有利于控制钢中硫化物形态。     

In-situ characterization of interface delamination by a new miniature mixed mode bending setup

A new miniature mixed mode bending (MMMB) setup for in-situ characterization of interface delamination in miniature multi-layer structures was designed and realized. This setup consists of a novel test configuration to accomplish the full range of mode mixities and was specially designed with sufficiently small dimensions to fit in the chamber of a scanning electron microscope (SEM) or under an optical microscope for detailed real-time fracture analysis during delamination. Special care was taken to minimize the effects of friction, the influence of gravity, and non-linearities due to the geometry of the setup. The performance of the setup was assessed using specially-designed test samples supported by finite element analyses. Delamination experiments conducted on homogeneous bilayer samples in mode I and mixed mode loading were visualized with a scanning electron microscope and showed the formation of small micro cracks ahead of the crack tip followed by crack bridging and a full crack, thereby demonstrating the advantages of in-situ testing to reveal the microscopic delamination mechanism.     

EMD与自适应滤波相结合的语音增强法

为了提高语音信号的信噪比,提出一种经验模态分解与自适应滤波相结合的语音增强法。对带噪语音进行经验模态分解,得到有限个固有模态函数,把所有的固有模态函数按顺序分成三组,将每一组所包含的固有模态函数叠加,得到三个子信号;对三个子信号进行自适应滤波,消除噪声;将降噪后的子信号重构得到增强后的语音。仿真实验表明,所提方法的语音增强效果优于自适应滤波。     

一种对汉语孤立字词进行声韵分割的时频方法

正确的声韵分离是汉语语音识别与合成等的基础和关键。针对传统声韵分割中,时域短时能量和过零率容易受到噪声干扰从而导致分割不准确的问题,结合语谱图所体现的时频信息对汉语孤立字进行了声韵分割,并进一步对信号进行经验模态分解和计算保号率,实现了一种对二字词的时频声韵分割方法。仿真实验结果显示,该方法对汉语孤立字和二字词的分割准确率分别达到了86.92%和77.47%。     

面向信号的仪器类驱动器设计与实现

针对自动测试系统（ATS）仪器可互换性的实现问题,借鉴IEEE1641信号与测试定义标准（Signal and Test Definition,STD）和IVI（Interchangeable Virtual Instrument）仪器类驱动思想,采用COM（Component Object Model）组件技术完成了基于信号接口的仪器类驱动器设计,同时实现了对不同种类开关的模型化描述和统一控制。该仪器驱动器在机载电子设备测试系统中得到实际的应用,实现了具有相同信号性能的同类仪器及不同仪器间的互换,减少了ATS升级维护的成本和时间。     

高速数据传输系统的WDM驱动程序的封装设计

高速数据传输系统多采用第三代高性能I/O总线——PCI Express总线作为实时数据传输接口,硬件接口一般需要接口芯片的支持从而使WDM驱动程序成为必需。以桥接芯片PEX8311为例设计了Windows XP系统环境下的WDM驱动程序,重点阐述了该驱动程序的数据传输和动态链接库（DLL）封装。封装后的驱动程序彻底解决了操作系统底层和应用层间交互调用的编程难题,其通用性强,只需在底层做少量芯片配置的修改便可移植作为新的接口芯片的驱动程序,而不需改动应用程序。     

The impact of EPI voxel size on SNR and BOLD sensitivity in the anterior medio-temporal lobe: a comparative group study of deactivation of the Default Mode

Objectives  To quantify the gain in time-series SNR that can be achieved in the amygdala by reducing EPI voxel size, and to assess the extent to which this advantage is carried through to statistical significance in a group fMRI study, using a cognitive task to trigger task-independent deactivation of anterior medial temporal structures. Materials and methods  Two groups of seven subjects were posed number-series tasks to induce deactivation of the Default Mode network. This is known from PET work to include the amygdala, which lies in a region of high magnetic field gradient. In 3 T imaging, one group was studied with high resolution EPI with 6 μl voxels, the other with lower resolution EPI with 17 μl voxels. Field maps were acquired to allow field gradients in relevant ROIs to be assessed. Results  Time-series SNR was 45% higher in the amygdala in the high resolution EPI data than in the low resolution data. In activation results, whilst there was good agreement between other areas, the involvement of the amygdala could only be demonstrated in the high resolution data. Conclusion  We find that reduction in signal dephasing afforded by high resolution EPI is realized as a substantial increase in SNR and BOLD sensitivity in group fMRI data. This has allowed the first demonstration of the involvement of the amygdala in the Default Mode in fMRI.     

MIMO非线性不确定系统二阶滑模控制

提出了一种基于有限状态机切换策略的多输入多输出二阶滑模控制算法。算法保证了传统滑模控制对参数变化和扰动不灵敏的特点,削弱了滑模控制的“抖动”现象。在上界未知的测量噪声和参数变化的情况下,算法通过滑模量及其微分的符号构成控制律,实现了系统的镇定。仿真结果表明算法在噪声环境下能保证系统的稳定性,对参数不确定具有较强的鲁棒性。算法结构简单,便于实现。