谐波法复杂周期信号周期测量

利用复杂周期信号频谱存在大量等间隔幅度起伏的谐波分量.提出了复杂周期信号的谐波法周期测量.谐波法之FT-Abs-IFT(FAIF)是周期信号频谱包络的反傅里叶变换,输出为频谱包络对应时域信号与周期冲击信号的卷积.其最大值是等幅度或起伏幅度(噪声条件下)的周期冲击信号,通过门限栓测和周期测量得到信号周期(或频率).最后通过与常用周期信号测量方法(同态滤波法、自相关法和AMDF)的仿真比较,表明对于复杂周期信号,谐波法之FAIF测量精度高,误差率低和良好的抗噪性能.     

一种应用于IEEE 802.11 a_n的高线性度InGaP/GaAs功率放大器

A three-stage 4.8-6 GHz monolithic power amplifier(PA) compatible with IEEE 802.11a/n designed based on an advanced 2μm InGaP/GaAs hetero-junction bipolar transistor(HBT) process is presented.The PA integrates input matching and closed-loop power control circuits on chip.Under 3.3 V DC bias,the amplifier achieves a ~31 dB small signal gain,excellent wide band input and output matching among overall 1.2 GHz bandwidth,and up to 24.5 dBm linear output power below EVM 3%with IEEE 802.11a 64QAM OFDM input signal.     

关于《检规》中错边量安全定级的讨论

就《检规》中对于容器错边量超标定级处理存在的问题提出了相应的解决办法,并采用有限元应力分析验证了该方法的合理性和可行性,最后提出了相关建议。     

Differential startle reactivity following central CCK-8S and systemic boc CCK-4 administration in mice: Antecedent stressor history and testing condition.

The influence of intraventricular cholecystokinin-8S (CCK-8S) and systemic N-t-Boc-Trp-Met-Asp-Pheamide (Boc CCK-4) was evaluated in the acoustic and fear-potentiated startle paradigms in CD-1 mice. In the light + tone startle condition, CCK-8S increased startle 168 hrs after administration, compared with saline. In the tone startle condition, CCK-8S decreased startle immediately and 24 hrs after administration, compared with saline. Among nonshocked mice, CCK-8S increased startle at 48 and 168 hrs, compared with saline. In the light + tone condition, 5 μg Boc-CCK-4 did not influence startle, whereas 15 μg Boc CCK-4 decreased startle immediately, 24 hrs, and 48 hrs following administration. Results demonstrate that antecedent environmental experiences interact with subsequent pharmacological challenges in provoking the temporal expression of alterations in startle magnitude. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

油气商品储量的经营概念剖析

以商品价值理论为基础，结合石油资源市场的实际情况，采用定性分析方法，就油气储量的经营概念进行了详细的剖析，其中包括油气储量的使用价值和价值，社会劳动和个别劳动以及油气储量价值在资源市场扔定性确定方法，在参与国际市场竞争中，不仅要研究各石油公司的社会必要劳动，其经营的重点应是减少个别劳动付出，即发挥成本优势。     

Factor analysis of category and magnitude scales of a technical attribute.

36 journeymen electronics maintenance personnel judged the complexity of 16 avionics circuits using the paired-comparison, magnitude-estimation, rank-order, and constant-sum procedures on 2 occasions. The basic scale values were standardized across the 16 circuits separately within each method and occasion. The standardized scale values were then intercorrelated and factor analyzed to test the hypothesis that 2 factors would account for the data. After considering 1-, 2-, 3-, and 4-factor solutions, a 2-factor solution was chosen as best fitting the data. These 2 factors suggested the taxonomy of "congnitive discrimination" and "contextual uncertainty" to account for Ss' scaling behavior. Most frequently, the paired-comparison and the constant-sum methods were most heavily loaded on Factor 1, "cognitive discrimination." The rank-order and the magnitude-estimation methods were most consistently loaded heaviest on "contextual uncertainty," Factor 2. The findings are interpreted in terms of their relationship to certain customary scaling classificatory schemes. (24 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

基于BP神经网络的地震预测研究

以地震特性的研究为出发点,以预报震级为最终目标,通过BP(Back Propagation)神经网络实现地震震级的预测算法,用仿真方法模拟个体行为对系统的影响,研究了地震震级的预测技术,并通过测试数据预测了系统的整体运行情况.     

基于支持向量机的地震预警震级快速估算研究

以更准确的估算地震预警(earthquake early warning,EEW)震级为目标,利用P波触发后3 s内的日本K-net强震数据,选取幅值参数、周期参数、能量参数、衍生参数这4大类共12个P波特征参数作为输入,构建基于支持向量机震级预测模型(support vector machine for earthquake magnitude estimation,SVM-M)。结果表明,比较传统的震级估算“τc方法”与“P d方法”,建立的SVM-M模型震级预测误差明显减小且不受震中距变化的影响,小震高估问题得到明显改善。2016年日本熊本地震主震(M j7.3)与2008年中国汶川地震主震(M s8.0)的震例分析结果表明,3 s时间窗不能匹配震源破裂全过程而出现了一定程度的震级低估,但仍可在P波触发后短时间窗内明确是大地震事件。建立的SVM-M模型可应用于地震预警震级快速估算。     

一种基于边缘图像的快速物体检测方法

为提高物体检测速度,提出一种利用边缘快速检测物体的方法。首先,求取图像中每一个像素的边缘模值和方向,依据模值定位边缘位置,依据位置和方向进行边缘分组;然后,通过计算边缘组的相似度来对候选的包围盒进行投票;最后,统计多尺度滑动窗口上各候选包围盒的投票得分,依据包围盒得分对包围盒进行过滤,再依据包围盒重合度对包围盒进行合并,并依据合并的包围盒数量判断其是否为物体,最终得到图像中各物体的包围盒。实验结果表明,相对于目前最新的物体检测算法相比,该方法的运算效率最高,且检索率和精确度高。     

基于Gabor纹理特征的人脸识别方法*

为了降低人脸Gabor特征的维数,提出了一种新的基于Gabor幅值的纹理表征(GMTR)方法用于人脸识别。GMTR由伽玛分布(ΓD)拟合Gabor幅值的分布来刻画,拟合的ΓD参数作为纹理特征。在FERET和Yale人脸库上的实验结果显示GMTR的识别性能优于传统的Gabor幅值特征,表明纹理特征具有更强的鉴别力。