Electrolyte infiltration in phosphazene-based dye-sensitized solar cells

We report here a study of phosphazene polymer and oligomer electrolyte infiltration into high surface area titanium dioxide electrodes and its effect on the performance of dye-sensitized solar cells. The effects of different cell assembly procedures on the electrochemical properties are examined, as well as the infiltration of electrolytes based on poly[bis(2-(2-methoxyethoxy)ethoxy)phosphazene] (MEEP), hexakis(2-(2-methoxyethoxy)ethoxy)cyclotriphosphazene (MEE trimer), and a linear short chain analogue into conventional titanium dioxide electrode mesoporous (nanosphere) films, microcolumns and nanowires. The effects of temperature, co-solvents, and the order of addition of the electroactive components are found to affect both the conductivity of the electrolytes and the electrochemical performance of the cells. Cross-sectional scanning electron microscopy (SEM) imaging is employed to examine the degree of electrolyte infiltration into the nanostructured electrodes as a function of filling conditions. Using these techniques, conditions are identified for achieving a high degree of pore filling by the three electrolyte systems. Increased power conversion efficiency is obtained when iodine is introduced after the heating and evacuation procedures required for maximum infiltration.     

广域覆盖场景中GSM与3G系统间干扰分析

结合山区覆盖特点,对GSM与WCDMA及CDMA1X系统间的干扰情况进行了确定性计算分析,给出了GSM与两系统共存与共站情况下的隔离要求与干扰规避方法。     

面向元余弦损失的少样本图像分类

目的 度量学习是少样本学习中一种简单且有效的方法,学习一个丰富、具有判别性和泛化性强的嵌入空间是度量学习方法实现优秀分类效果的关键。本文从样本自身的特征以及特征在嵌入空间中的分布出发,结合全局与局部数据增强实现了一种元余弦损失的少样本图像分类方法(a meta-cosine loss for few-shot image classification,AMCL-FSIC)。方法 首先,从数据自身特征出发,将全局与局部的数据增广方法结合起来,利于局部信息提供更具区别性和迁移性的信息,使训练模型更多关注图像的前景信息。同时,利用注意力机制结合全局与局部特征,以得到更丰富更具判别性的特征。其次,从样本特征在嵌入空间中的分布出发,提出一种元余弦损失(meta-cosine loss,MCL)函数,优化少样本图像分类模型。使用样本与类原型间相似性的差调整不同类的原型,扩大类间距,使模型测试新任务时类间距更加明显,提升模型的泛化能力。结果 分别在5个少样本经典数据集上进行了实验对比,在FC100(Few-shot Cifar100)和CUB(Caltech-UCSD Birds-200-2011)数据集上,本文方法均达到了目前最优分类效果;在MiniImageNet、TieredImageNet和Cifar100数据集上与对比模型的结果相当。同时,在MiniImageNet,CUB和Cifar100数据集上进行对比实验以验证MCL的有效性,结果证明提出的MCL提升了余弦分类器的分类效果。结论 本文方法能充分提取少样本图像分类任务中的图像特征,有效提升度量学习在少样本图像分类中的准确率。     

移动网络中MCL-MDS定位算法

在深入了解移动网络中蒙特卡洛(MCL)算法的基础上,针对MCL算法在锚节点率较少时定位精度较差的缺点提出了一种结合MCL算法与分布式MDS-MAP(Distribute MDS-MAP,DMDS-MAP)算法的自适应算法MCL-MDS定位算法。算法首先使用所设计的DMDS-MAP算法对节点进行定位,随后利用MCL算法中的过滤原理对定位结果进行验证,若通过验证,则采用该定位结果,否则采用MCL算法的定位结果。仿真显示,MCL-MDS定位算法能在各种环境下获得较好的定位效果。     

新的室内移动机器人自定位方法

针对现有室内移动机器人自定位方法中存在的定位精度不高,随时间积累定位误差增大,复杂室内环境下信号存在多径效应和非视距效应等问题,提出了一种基于蒙特卡罗定位(MCL)的新的移动机器人自定位方法。首先,通过分析基于无线射频识别(RFID)技术的移动机器人自定位系统,建立机器人运动模型;然后,通过分析基于接收信号强度指示(RSSI)的移动机器人自定位系统,提出机器人移动过程的观测模型;最后,针对粒子滤波定位执行效率不高的问题,提出粒子剔除策略和依据粒子方位赋予粒子权值策略,提高系统的定位精度和执行效率。仿真实验表明,机器人在移动过程中的自定位误差在X轴和Y轴方向上为3 cm,传统定位算法误差为6cm,新算法定位精度提高近1倍,且算法具有很好的鲁棒性。     

Bioaccumulation of metals in timber and edible fruit trees growing on reclaimed coal mine overburden dumps

A field study was carried out (IB-valley areas, Mahanadi Coalfields Limited) to assess the bioaccumulation of heavy metals in the tree species, changes in physicochemical and nutritional properties of mine soils, and bioavailability of metals. Only limited amount of topsoil was added in the plantation pits in the ratio of 1:4 (topsoil: weathered mine spoils), which significantly improved pH and fertility. Out of the six analysed metals, the highest concentration was found for Fe followed by Mn > Zn > Cu > Cd > Ni in mine soils. Metal accumulation and bioconcentration factor (BCF) for each metal were calculated for edible fruit (Mangifera indica, Anacardium occidentale, Psidium guajava, Artocarpus heterophyllus and Syzygium cumini) and timber trees (Acacia mangium, Eucalyptus spp., Techtona grandis and Gravellia robusta). Higher metal accumulations were observed in edible fruit trees than control plants growing in garden soil. Significant variations of metal accumulations were observed amongst timber tree species with higher BCF values. Present study suggested that in spite of addition of topsoil, no significant improvement was observed for the bioavailable metal concentration in the mine soils. Additional care should be undertaken, if edible fruit trees are considered for afforestation of coal mine spoils.     

基于MC L与 KN N的混合聚类算法

MCL是一种图聚类算法,针对MCL计算过程会产生小聚类及边缘节点从团中脱离出来的问题,提出了一种基于MCL与KNN相结合的混合聚类算法。该算法利用KNN的分类特点,以MCL聚类得到的聚类表为依据,通过KN N对小聚类中的元素进行再分类,以提高聚类的质量。实验证明此方法是可行的,改进后的算法能使聚类质量有所提高。     

使用NDT激光扫描匹配的移动机器人定位方法

提出一种将基于扫描匹配的蒙特卡洛定位方法,作为移动机器人完成自主任务的鲁棒性定位方法． 采用一种新的正态分布转换(NDT)激光扫描匹配算法,将从单个激光扫描重构的2D离散数据点集转换成2维平面内分段连续可微的概率分布,并使用Hessian矩阵法与另外的扫描相匹配,可以避免点与点之间对应的复杂问题．实验结果表明,该定位算法可以利用自然环境特征有效地完成室内环境下的自主定位．     

Magnetic manipulation device for the optimization of cell processing conditions

Variability in human cell phenotypes make it's advancements in optimized cell processing necessary for personalized cell therapy. Here we propose a strategy of palm-top sized device to assist physically manipulating cells for optimizing cell preparations. For the design of such a device, we combined two conventional approaches: multi-well plate formatting and magnetic cell handling using magnetite cationic liposomes (MCLs). From our previous works, we showed the labeling applications of MCL on adhesive cells for various tissue engineering approaches. To feasibly transfer cells in multi-well plate, we here evaluated the magnetic response of MCL-labeled suspension type cells. The cell handling performance of Jurkat cells proved to be faster and more robust compared to MACS (Magnetic Cell Sorting) bead methods. To further confirm our strategy, prototype palm-top sized device “magnetic manipulation device (MMD)” was designed. In the device, the actual cell transportation efficacy of Jurkat cells was satisfying. Moreover, as a model of the most distributed clinical cell processing, primary peripheral blood mononuclear cells (PBMCs) from different volunteers were evaluated. By MMD, individual PBMCs indicated to have optimum Interleukin-2 (IL-2) concentrations for the expansion. Such huge differences of individual cells indicated that MMD, our proposing efficient and self-contained support tool, could assist the feasible and cost-effective optimization of cell processing in clinical facilities.     

WCDMA与PHS相互干扰的MCL法分析

目前中国电信和中国网通的PHS网络已经具有一定的规模，在全国多数城市PHS网络已经成为继中国移动和中国联通GSM网络之后的第三大无线网络。然而，由于PHS 1900MHz和WCDMA所用频率非常靠近，随着3G脚步的不断接近，研究PHS 1900MHz和WCDMA相互干扰显得刻不容缓。现通过MCL法对它们之间的干扰进行了研究分析，得出了一些对3C网络建设具有指导意义的结论。