一种灵敏度准则矩阵构成的新方法

针对变形监测网的特点,对准则矩阵加以研究,提出了一种构造灵敏度准则的新方法。用该方法通过典型模型下的实例进行二次优化,论证了该方法的可行性和合理性。     

融合深度学习与最大间距准则的人脸识别方法

当前,人脸识别技术遇到的突出问题是光照、姿态、遮挡和表情等因素所引起的识别精度的下降,这些问题是人脸识别系统不完美的主要原因,深度学习是一种新的方法,可有效解决这些问题。首先通过引入深度学习算法进行多层次的学习,然后提取高层特征进行人脸描述,最后应用最大间距准则减小最小二乘估计产生的重建误差,实现有效的面部识别分类。该算法在ORL、CAS-PEAL和扩展Yale-B人脸数据库中进行了不同光照、姿态、遮挡、表情和容貌特征变化条件下的仿真实验。结果表明,所提出的算法比传统线性分类算法具有更高的效率和准确度。     

一种新的基于最小二乘准则的超宽带多用户检测算法

本文通过由跳时码唯一确定的编码矩阵将超宽带无线多址通信信道模型建模为一种线性模型,提出了一种新的基于最小二乘准则的超宽带多用户检测算法。该算法把最小二乘准则公式： 写成矩阵形式并且利用矩阵Moore-Penrose逆递推公式实现的自适应递推算法直接对编码矩阵进行更新,从而得到了一种新的超宽带多址通信自适应多用户检测算法。计算机模拟结果表明,该算法比传统的RLS算法有更好的收敛速度,并且相应的自适应多用户检测器比传统的检测器有更好的检测性能。     

Geophysical criterion of pre-outburst coal outsqueezing from the face space into the working

The initial process of coal and gas sudden outburst is studied in the article when under the influence of rock and gas pressure the part of a coal seam layer(a coal section) is squeezed out from the mouth of the future outburst cavity in a jump-like manner into the working. Geo-mechanical criterion for a part of a coal seam layer outsqueezing in the form of the relation of active(squeezing out) and passive(preventing the outsqueezing) forces is defined in the article. Based on it, the geophysical criterion is defined by expressing basic physical parameters through geo-physical ones: the current stress is defined by spectral-acoustic method through the ratio of high frequency and low-frequency components of an acoustic signal, which is generated into a face working space by the mining equipment operating in the face; in-situ gas pressure is defined by gas analytical method by the concentration of methane in the atmosphere of the working; the strength of the most broken coal layer is defined by a strength measuring device(a device for measuring the depth of a steal cone punched into the coal by a spring mechanism). This paper studies the influence of acoustic, strength and filtrating and collecting properties of a face working space on the limit value of an obtained geophysical criterion of pre-outburst squeezing of a coal ‘‘plug" out of the mouth of the future outburst cavity into the working.     

类人猿型机器人前向四足步行的研究

类人猿机器人是一种具有多种移动方式的新型机器人。同常规的四足步行机器人相比,该机器人的结构具有特殊性和复杂性。为了实现拥有这种结构机器人的前向四足步行,我们提出了新的步态、轨迹规划、稳定平衡判据。仿真和实验结果证实了这些分析的合理性。     

显微镜自动聚焦算法的研究

自动聚焦技术是全自动控制显微镜系统中的核心部分。自动聚焦算法综合了快速灰度差分法和高精度拉普拉斯函数法的优点,根据聚焦过程中各阶段聚焦评价函数的梯度选择合适的焦距评价函数和步长,并通过改变步长实现了从快速粗调到精确细调的过渡。实验表明,与传统的单一聚焦评价函数和单一步长的聚焦算法相比较,该算法在保证聚焦速度的情况下显著地提高了聚焦位置的精度。     

注册建筑师制度在中国的建立及管理统计

文章重点从政策层面比较系统地回顾了中国建设行业注册建筑师执业管理制度从建立到发展演变的过程以及注册资格国际互认方面的情况。     

Damage assessment on plate-like structures using a global-local optimization approach

A new method for finite element model updating using simulated data is presented. A global-local optimization (GLO) approach was adopted to adjust the uncertain properties of the FE model by minimizing iteratively the differences between the measured dynamic modal parameters and the corresponding analytical predictions. In contrast with most of the existing updating techniques, which minimize modal force errors, objective functions based on Coordinate Modal Assurance Criterion (COMAC) and Frequency Response Assurance Criterion (FRAC) were employed. The GLO procedure was employed to minimize the norm of these vectors by updating the physical model variables (thickness, Young modulus, etc.). The proposed model updating procedure was applied to damage localization and quantification of structures, whose damage characteristics can be represented by a reduction of the element bending and axial stiffness. Results showed that a significant reduction in terms of computer run-time and improved damage assessment can be achieved. The procedure is illustrated on a plate-like structure by measuring dynamic properties before and after structural changes for four different damage cases.     

Fisher极小准则不相关空间算法及其在人脸识别中的应用

不相关空间算法是一种基于广义Fisher准则求解不相关鉴别矢量集的快速算法,但该算法要求总体散度矩阵可逆。针对高维小样本的情况,文中提出求解不相关鉴别矢量集的改进方法。该方法的基本思路是在类间散度矩阵的值空间中运用广义Fisher极小准则求解鉴别矢量集,并讨论在该子空间中进行求解的合理性。针对高维情况下类间散度矩阵值空间的计算效率问题,提出首先利用PCA算法将数据降维,然后在低维空间中求解值空间的策略并讨论其合理性。在ORL人脸数据库上的实验验证该方法的有效性,其识别率高于传统的Fisher脸方法和不相关空间算法。     

基于PCA扩展的判别性特征融合

提出两个判别性的特征融合方法——主成分判别性分析和核主成分判别性分析。基于主成份分析和最大间隔准则理论,构造一个多目标规划模型作为特征融合的目标。随后,该模型被转化成一个单目标规划问题并通过特征分解的方法求解。此外,将一个近似分块对角核矩阵K分成c(c为数据集中的类别数)个小矩阵,并求出它们的特征值和特征向量。在此基础上,通过向量代数处理得到一个映射矩阵α,当核矩阵K投影到α上,同类样本的相似信息能最大程度地得到保持。本文中的实验证实两种方法的有效性。