动态核的形态分解联想算法及抗随机噪声研究

在灰度图像分解算法和动态核形态联想记忆网络的基础上,提出了一种新的联想记忆算法--动态核的形态分解联想算法.该方法显著地提高了联想记忆抗随机噪声的能力,较好地解决了灰度图像在含噪时的联想记忆和识别的问题,从而给出了一种恢复含噪灰度图像的途径,并把该方法推广到了彩色图像的处理.通过实验,验证了该方法的良好性能,取得了理想的结果.     

新型推挽激励的谐振式微机械静电场传感器

为了提高传感器的信噪比(signal-to-noise ratio,SNR),降低器件的耦合干扰,基于表面微机械工艺设计制作了一种新型推挽激励方案的谐振式微型静电场传感器.该传感器敏感结构主要包括静电梳齿激励电极、屏蔽电极和感应电极3部分.工作时,在该传感器敏感结构同一侧梳齿的固定端上加入反相相等的激励信号,实现差分激励,使屏蔽电极在谐振频率处振动,从而在感应电极上感应出差分电流信号,通过相关检测方式检测出传感器该输出信号的二次谐波分量,即可获得被测静电场强度大小.该传感器的新型结构设计和激励检测方案可获得最大电场响应灵敏度,并有效抑制了同频耦合串扰和共模噪声.常压室温(约25℃)测试结果表明,在量程(-15 kV/m～15 kV/m)内,只需要5 Vpp激励电压,其中线性度(端基线性度)达到了1.03%,分辨率优于200 V/m.     

Pulsed Electromagnetic Field (PEMF) Treatment Reduces Lipopolysaccharide-Induced Septic Shock in Mice

Despite advances in medicine, mortality due to sepsis has not decreased. Pulsed electromagnetic field (PEMF) therapy is emerging as an alternative treatment in many inflammation-related diseases. However, there are few studies on the application of PEMF therapy to sepsis. In the current study, we examined the effect of PEMF therapy on a mouse model of lipopolysaccharide (LPS)-induced septic shock. Mice injected with LPS and treated with PEMF showed higher survival rates compared with the LPS group. The increased survival was correlated with decreased levels of pro-inflammatory cytokine mRNA expression and lower serum nitric oxide levels and nitric oxide synthase 2 mRNA expression in the liver compared with the LPS group. In the PEMF + LPS group, there was less organ damage in the liver, lungs, spleen, and kidneys compared to the LPS group. To identify potential gene targets of PEMF treatment, microarray analysis was performed, and the results showed that 136 genes were up-regulated, and 267 genes were down-regulated in the PEMF + LPS group compared to the LPS group. These results suggest that PEMF treatment can dramatically decrease septic shock through the reduction of pro-inflammatory cytokine gene expression. In a clinical setting, PEMF may provide a beneficial effect for patients with bacteria-induced sepsis and reduce septic shock-induced mortality.     

Structural control of Fe-based alloys through diffusional solid/solid phase transformations in a high magnetic field

Abstract

A magnetic field has a remarkable influence on solid/solid phase transformations and it can be used to control the structure and function of materials during phase transformations. The effects of magnetic fields on diffusional solid/solid phase transformations, mainly from austenite to ferrite, in Fe-based alloys are reviewed. The effects of magnetic fields on the transformation temperature and phase diagram are explained thermodynamically, and the transformation behavior and transformed structures in magnetic fields are discussed.     

Reliability Modeling for Ultrathin Gate Oxides Subject to Logistic Degradation Processes with Random Onset Time

With device dimension downscaling, the oxide thickness reduction less than 3 nm in a metal–oxide–semiconductor structure has led to important changes in degradation mechanisms and failure modes. After the first breakdown event of ultrathin gate oxides, the leakage current presents a less dramatic and noisy continuous breakdown mode called progressive breakdown. In this article, we characterize the overall breakdown process of ultrathin gate oxides as a randomized logistic degradation process with a random onset time. The explicit result of the lifetime distribution is derived on the basis of this logistic degradation model. A numerical example is provided to calculate the lifetime distribution. The simulated lifetime data of our model fits lognormal distribution better than the Weibull distribution, which agrees with the experimental work in literature. Copyright © 2012 John Wiley & Sons, Ltd.     

一种在线学习的目标跟踪与检测方法

针对现有在线学习跟踪方法缺乏监督机制的缺点,提出一种新的跟踪框架。以随机森林在线学习理论构造分类器作为目标检测器,用SURF特征点匹配方法作为目标跟踪器,跟踪过程中用可靠跟踪的结果形成对检测结果的监督机制,不对低置信度样本或错误样本进行学习,避免了分类器分类精度的逐渐下降,同时在跟踪失败时用目标检测器对目标进行重新捕获,形成跟踪、学习、检测三者有机结合的跟踪框架。对不同视频序列的测试结果表明,本文算法能有效避免目标出现较大外观变化或被大面积遮挡等复杂情况下的跟踪失败问题。     

非对称电阻场板场效应器件击穿特性仿真分析

利用TCAD仿真研究一种二维紧耦合电阻场板电流调制原理下的物理模型与最优化结构。通过优化关键工艺与材料参数,改善器件漂移区尖峰电场,最终在相同漂移区掺杂下击穿电压较一维PN结理论击穿电压提升273%,相同归一化击穿电压10%变化范围下,漂移区电荷变化允许冗余范围比现有传统PN超结拓宽15倍。相较于对称电阻场板场效应器件,在现有工艺下非对称优化电阻场板场效应器件能够更好的实现结构小型化与高密度的设计。     

一种沟槽-场限环复合终端结构的设计

为了改善硅功率器件击穿电压性能以及改善IGBT电流的流动方向,提出了一种沟槽-场限环复合终端结构。分别在主结处引入浮空多晶硅沟槽,在场限环的左侧引入带介质的沟槽,沟槽右侧与场限环左侧横向扩展界面刚好交接。结果表明,这一结构改善了IGBT主结电流丝分布,将一部分电流路径改为纵向流动,改变了碰撞电离路径,在提高主结电势的同时也提高器件终端结构的可靠性;带介质槽的场限环结构进一步缩短了终端长度,其横纵耗尽比为3.79,较传统设计的场限环结构横纵耗尽比减少了1.48%,硅片利用率提高,进而减小芯片面积,节约制造成本。此方法在场限环终端设计中非常有效。     

一种改进的ALOHA防碰撞算法

针对现有动态帧时隙ALOHA防碰撞算法在射频识别系统中的标签识别效率最高只能达36.8%的问题,利用标签码元序列的唯一性,改进经典的动态帧时隙ALOHA防碰撞算法中随机选择时隙的方式,提出一种基于标签码元的碰撞序列进行时隙选择的方法,有效降低了标签碰撞的概率,从而提高系统识别效率。仿真结果表明改进的ALOHA标签防碰撞算法识别效率最低为37.5%,随着实际标签数目与碰撞位序列所能确定的标签数目越接近,识别效率越高,最高能达到100%,明显优于现有的动态帧时隙ALOHA算法。