Statistical Inference in a Linear Model for Spatially Located Sensors and Random Input

In the paper we consider a random linear model for observations provided by spatially located sensors measuring signals coming from one source. For this model a set of sufficient and complete statistics are found, and it is shown that the maximum likelihood estimators of unknown parameters (characteristics of the source) are functions of those statistics. The problem of nonnegative estimators of variance components of the model is shortly discussed. Comparisons of the mean squared errors of several estimators are given. Numerical example concerning hunting for defects in solar cells is considered in details.     

Pixel-based partially coherent image method for maskless lithography system

Maskless lithography (ML) provides a fast and low-cost method for projecting the images of IC or micro features onto photoresist. However, it needs an efficient simulation method to evaluate the performance of lithography process. In this paper, a pixel-based partially coherent image method for digital micro-mirror device (DMD) based ML is proposed based on the linear invariant theory. In our method, the mask is sampled by DMD pixel (each pixel corresponding to each micro-mirror) and expressed by rect function. Using the shift theory of Fourier transform and the stacked pupil operator approach, we build a matrix Φ for system response function of rect function. If the DMD pixel state matrix is S, then the aerial image can be calculated with two matrix multiplication I(x,y) = SΦ.     

Charge transfer of 1 ML C60/Ag（100）

We have measured the synchrotron radiation photoelectron spectra of monolayer C60on Ag（100）, By calculating the intensity ratios between the LUMO bands and the two deeper bands （HOMO and HOMO- 1）, we estimate the amount of the charges transferred from Ag（100） to C60 within the range of 1 e to 1.8 e. The results dismay the expectation of surface superconductivity and afford a good reference for further studies of the monolayer C60/Ag（100） system.     

机器学习筛选用于气体吸附分离和存储的金属有机骨架材料

金属有机框架（Metal-organic framework ,MOF）因其高孔隙率、高比表面积和结构可调性,在气体吸附分离领域广泛应用。随着MOF数量激增,传统分子模拟和实验方法验证MOF性能成本高且速度慢,因此目前MOF筛选工作已转向高通量计算辅助的机器学习（Machine-learning,ML）。机器学习作为一种高效的大数据处理方法,能够在高通量筛选（High-Throughput Computational Screening,HTCS）的基础上对数据进行拟合,从而快速而准确地筛选出气体吸附分离材料,并深入挖掘其结构与性能之间的关系。本文回顾了近年机器学习应用于MOF筛选的研究。本文重点讨论了一些运用机器学习从大量结构中筛选出可用于CH4、H2和CO2等气体吸附分离与储存的MOF材料的工作。同时,我们梳理了当前MOF材料筛选工作中的研究思路和进展,并指出了机器学习在筛选MOF材料工作中面临的一些瓶颈和挑战。最后,对该领域的未来发展前景进行了展望。     

Scanning tunneling microscopy study of fullerenes

Scanning tunneling microscopy investigations of adsorption and film growth of various fullerenes on semiconductor and metal surfaces are reviewed. The fullerenes being studied are C₆₀, C₇₀, C₈₄, Sc@C₈₂ and Y@C₈₂ and the substrates being used for adsorption are Si (111), Si (100), Ge (111), GaAs (110), GaAs (001), Au (111), Au (110), Au (100), Cu (111) and Ag (111) surfaces.     

软件锁相环在惯性动量轮转速控制中的应用研究

研究了基于DSP和软件锁相环算法的惯性动量轮用高速(最高转速30000r／min)高精度转速控制系统。首先分析了组成软件锁相环的三个环节，设计了软件锁相环算法；然后与高精度反电势换相模块ML4425相结合，设计并实现了一种高精度无刷直流电机数字控制系统。该系统能在任意转速下进行稳速控制，且当转速高于3000r／min时，能实现1％％的转速控制精度。     

无重复因析试验中散度效应的ML估计

在无重复因析试验中,若因子A,B的散度效应显著,则不论其交互效应AB的散度效应是否显著,其散度效应的现有估计常常是有偏的,从而导致其被错误地识别为显著效应.提出了散度效应的一种新的估计方法(称为ML估计),并给出了ML估计的方差的精确表达形式,证明了在一类模型中,交互效应AB的散度效应的ML估计是无偏的.最后,将ML估计与现有的常用估计进行了比较.     

用菲涅尔平板天线实现相干来波方向的估计

菲涅尔平板天线由于其具有宽频带、重量轻、馈电简单和易于共形等优点，可替代透镜天线用于毫米波成像或用作微波移信通信中的接收天线单元等。同时，其作为聚焦器件，通过与馈源阵列组合，可以构成一种新型的自适应天线应用于蜂窝移动通信的基站智能天线等领域，基于其良好的工程应用前景，本文在对有关研究工作简要回顾的基础上，深入分析其去相关机理，应用极大似然估计算法实现了相干信号源在大角度入射的情形下，菲涅尔平板天线对来波方向的有效估计，避免了应用MUSIC算法进行来波方向估计所遇到的观察区域的限制问题，仿真计算和有关文献数值分析结果的比较表明，菲涅尔平板天线具有良好的空间谱估计性能。