Novel delay-distribution-dependent stability analysis for continuous-time recurrent neural networks with stochastic delay

<正>In this paper,the problem of delay-distribution-dependent stability is investigated for continuous-time recurrent neural networks(CRNNs) with stochastic delay.Different from the common assumptions on time delays,it is assumed that the probability distribution of the delay taking values in some intervals is known a priori.By making full use of the information concerning the probability distribution of the delay and by using a tighter bounding technique(the reciprocally convex combination method),less conservative asymptotic mean-square stable sufficient conditions are derived in terms of linear matrix inequalities(LMIs).Two numerical examples show that our results are better than the existing ones.     

基于昆虫复眼技术的全视场运动目标监测系统

以昆虫复眼的仿生技术为基础,应用线阵CCD图像传感器、CPLD、AVR微处理器及运动目标检测算法实现了一种全视场监测系统,能简单而精确地探测出运动物体的角坐标和角运动速度。不仅对系统的硬件设计和软件设计进行了讨论,而且对应用的算法也进行了的介绍。     

Active noise attenuation in ventilation windows

The feasibility of applying active noise control techniques to attenuate low frequency noise transmission through a natural ventilation window into a room is investigated analytically and experimentally. The window system is constructed by staggering the opening sashes of a spaced double glazing window to allow ventilation and natural light. An analytical model based on the modal expansion method is developed to calculate the low frequency sound field inside the window and the room and to be used in the active noise control simulations. The effectiveness of the proposed analytical model is validated by using the finite element method. The performance of the active control system for a window with different source and receiver configurations are compared, and it is found that the numerical and experimental results are in good agreement and the best result is achieved when the secondary sources are placed in the center at the bottom of the staggered window. The extra attenuation at the observation points in the optimized window system is almost equivalent to the noise reduction at the error sensor and the frequency range of effective control is up to 390 Hz in the case of a single channel active noise control system.     

集成成像三维显示系统显示性能的研究进展

集成成像三维显示技术是利用透镜阵列获取和显示立体图像的一种三维显示技术。本文首先综述了集成成像三维显示系统的特点,考虑其系统性能主要受分辨率、景深和观看视场角的限制,对近年来集成成像三维显示系统在增大分辨率、景深和观看视场角方面的研究进展做了综合论述,比较分析了各种改进方法的优劣。最后,对我国集成成像三维显示技术的研究现状进行了总结,并简述了本研究小组在该领域取得的若干研究成果。     

应用可见/近红外光谱技术快速检测果珍中二氧化钛的含量

为了实现快速检测果珍中的二氧化钛含量,提出了应用近红外光谱技术结合化学计量学的快速检测方法。研究采用了320份果珍样本进行光谱特性的检测,其中200个样本用来建模,120个样本进行预测。首先比较了标准正态变量校正(SNV)、变量标准化(Normalize)、多元散射校正(MSC)等6种不同的数据预处理方法对偏最小二乘法(PLS)建模预测效果的影响。然后将PLS模型与应用主成分(PC)建立的主成分-神经网络校正(PC-ANN)模型进行比较。结果表明,MSC预处理的效果最好,PLS模型的最佳主成分数为7,预测值与标准值的相关系数R2达0.900 8,预测标准误差RMSEP为0.05。PC-ANN模型预测值与标准值的R2为0.868 4,RMSEP为0.04。说明PLS模型比PC-ANN模型的预测效果好。同时本研究也说明能够应用可见/近红外技术对二氧化钛进行快速定量测定。     

聚酰胺酸合成工艺研究

 采用超声波在线测量溶液粘度的方法，研究了合成聚酰胺酸过程中实验条件的影响，并对实验结果进行分析，由此确定了合成高分子量聚酰胺酸的最佳实验条件。研究表明：在加料次序为先加二胺后加二酐（二酐与二胺的摩尔比为1.01～1.02：1）、试剂中含水量尽可能少，反应温度0～5℃、反应时间以溶液粘度到达最大值为止的条件下，所合成的聚酰胺酸溶液粘度最大，可满足惯性约束聚变（ICF）充气腔靶端口膜的需要。     

Robust Preparation of Atomic Concatenated Greenberger–Horne–Zeilinger States via Shortcuts to Adiabaticity

Here, a protocol for robust preparation of an atomic concatenated Greenberger–Horne–Zeilinger (C‐GHZ) state via shortcuts to adiabaticity (STA) is proposed. The devices for implementing the protocol consist of atoms, cavities, and the optical fibers, which are feasible with current technology. The atoms are trapped in the separated cavities allowing individual control over each atom with classical fields. STA helps to design Rabi frequencies of classical fields so that the atoms can be driven from the initial states to the target states. The numerical simulations show that the protocol holds robustness against atomic spontaneous emissions and photonic leakages. Thus, the protocol may be realized by experiments in the near future.     

高光谱成像技术的油菜叶片氮含量及分布快速检测

应用高光谱成像技术实现了油菜苗-花-角果整个生命期叶片氮含量的快速检测和氮素水平分布的可视化。采集三个生长时期共计420个叶片样本的高光谱图像信息(380～1 030 nm),提取图像中感兴趣区域的平均光谱数据,经过不同光谱预处理后,利用连续投影算法(SPA)选择特征波长,将提取的12个特征波长(467,557,665,686,706,752,874,879,886,900,978和995 nm)作为自变量,叶片氮含量作为因变量,分别建立偏最小二乘法(PLS)和最小二乘-支持向量机(LS-SVM)模型。SPA-PLS和SPA-LS-SVM模型对叶片氮含量的预测相关系数R_P分别为0.807和0.836,预测均方根误差RMSEP分别为0.387和0.358。高光谱图像中的每一个像素点都有对应的光谱反射值,利用结构简单、更易提取回归系数的SPA-PLS模型,快速计算出12个特征波长下高光谱图像中每个像素点对应的氮含量预测值,结合像素点的空间位置生成氮素浓度的叶面分布图。可视化分布图详细且直观的反应出同一叶片内部或不同叶片之间氮含量的差异。结果表明,应用高光谱成像技术分析整个油菜生长期的叶片氮含量及其可视化分布是可行的。     

Preparation of nano-hydroxyapatite/poly(<Emphasis Type="SmallCaps">l</Emphasis>-lactide) biocomposite microspheres

Nano-hydroxyapatite (HA)/poly(l-lactide) (PLLA) composite microspheres with relatively uniform size distribution were prepared by a solid-in-oil-in-water (s/o/w) emusion solvent evaporation method. The encapsulation of the HA nanopaticles in microshperes was significantly improved by grafting PLLA on the surface of the HA nanoparticles (p-HA) during emulsion process. This procedure gave a possibility to obtain p-HA/PLLA composite microspheres with uniform morphology and the encapsulated p-HA nanoparticle loading reached up to 40 wt% (33 wt% of pure HA) in the p-HA/PLLA composite microspheres. The microstructure of composite microspheres from core-shell to single phase changed with the variation of p-HA to PLLA ratios. p-HA/PLLA composite microspheres with the diameter range of 2–3 μm were obtained. The entrapment efficiency of p-HA in microspheres could high up to 90 wt% and that of HA was only 13 wt%. Surface and bulk characterizations of the composite microspheres were performed by measurements such as wide angle X-ray diffraction (WAXD), thermal gravimetric analysis (TGA), environmental scanning electron microscope (ESEM) and transmission electron microscopy (TEM).     

基于ZY-3 CCD相机数据的暗像元大气校正方法分析与评价

ZY-3是我国首颗民用高空间分辨率光学传输型立体测图卫星, 可为国土资源调查、生态环境监测等发挥重要作用, 而大气校正是制约其广泛定量应用的关键问题之一。由于实测地面光谱数据和大气参数难以实时获取, 针对这种情况下如何反演得到高空间分辨率卫星精确的地表反射率这一问题, 基于2012年内蒙古野外实验实测数据, 对四种暗像元大气校正方法进行了分析与评价研究。分析了四种暗像元大气校正算法中的关键参数对ZY-3 CCD数据应用效果的影响, 结果表明: (1)四种暗像元大气校正方法在第1,2和3波段均有明显的校正效果, 其中DOS4方法在第4波段大气校正效果最好, DOS1和DOS3方法在第4波段大气校正效果不明显, DOS2方法在第4波段大气校正效果最差。(2)DOS1方法大气校正结果在4个波段的相对误差均大于10%。DOS2方法在第1波段校正效果最好(AE=0.001 9和RE=4.32%), 而在第4波段校正误差最大(AE=0.0464和RE=19.12%)。DOS3方法大气校正结果在4个波段的相对误差均约10%左右。(3)DOS4方法大气校正结果在4个波段的绝对误差均小于0.02和相对误差均小于10%, 大气校正精度最高。