Dense Y-doped ion conducting perovskite films of BaZrO3, BaSnO3, and BaCeO3 for SOFC applications produced by powder aerosol deposition at room temperature

State-of-the-art solid oxide fuel cells (SOFC) are based on oxide ion conducting zirconia electrolytes, typically doped by yttria or scandia. Major drawback of these systems are their high operation temperatures of 800 °C and above. These are necessary for a sufficient ionic conductivity. Instead of oxide ions, also protonic charge carriers could be used in SOFC. The material classes of barium zirconates (BaZrO₃), barium stannates (BaSnO₃), and barium cerates (BaCeO₃) are described as good proton conductors, especially when the B-site of the ABO₃ perovskite structure is aliovalently doped by yttria. Their protonic conductivity values in the moderate temperature regime up to 600 °C are comparable to YSZ at 800 °C, making these compounds ideal candidates for a usage in future SOFC. Unfortunately, very high sintering temperatures up to 1800 °C are required to process dense and therefore gas-tight solid electrolyte membranes. However, a novel spray coating method called powder aerosol deposition (PAD, also known as AD) enables to form fully dense ceramic films directly at room temperature without any necessary sintering processes. Films are deposited from ceramic powders that are accelerated by a dry carrier gas flow under vacuum conditions.In this work, we investigated the film formation of three different barium based perovskite ceramics, namely yttria doped barium zirconate, barium stannate, and barium cerate by powder aerosol deposition. The optical and mechanical quality of films was evaluated using scanning electron microscopy and microhardness indentation and their crystallographic properties were characterized by X-ray diffraction. The electrical behavior was analyzed by electrochemical impedance spectroscopy and DC polarization methods up to temperatures of 1000 °C and 800 °C, respectively. Furthermore, a preliminary study about the film formation on porous electrodes was conducted.     

我国红柱石选矿工艺与分选理论的研究进展

介绍了红柱石的性质及其用途,从选矿理论与选矿工艺的角度,总结了我国红柱石选矿工艺和分选理论的现状及进展,对红柱石选矿技术发展的趋势进行了归纳和总结。     

安全疏散基础参数的量化

采用实测统计法对人流密度、疏散速度和通行流量3个基础参数及其相互关系进行研究,分析参数特性及其在疏散设计中的应用,建立基础参数的关系模型。通过参数模型分析认为合理的疏散设计应当控制人流流量接近但不超过通道的峰值流量,保证水平通道内的人员密度不大于1.6人/m2的极限值,楼梯内的人流密度不大于2.5人/m2极限值。     

基于多输入密集连接神经网络的遥感图像时空融合算法

为了解决地表反射率遥感卫星Landsat和MODIS影像的时空融合问题,文中提出基于多输入密集连接网络的遥感图像时空融合算法.首先提出多输入的密集连接网络,学习包含连续时刻间差异信息的过渡遥感影像.基于差异相似假设,融合网络学习得到的2幅过渡影像与已知的2幅高空间分辨率影像,得到最终的预测影像.对Landsat遥感影像和MODIS遥感影像的融合实验表明,文中算法在各项定量指标中均较优,最终的预测图像也可表明,文中算法对噪声具有较好的鲁棒性,能较好地恢复细节信息.     

基于深度Q网络的人群疏散机器人运动规划算法

针对公共场合密集人群在紧急情况下疏散的危险性和效果不理想的问题，提出一种基于深度Q网络（DQN）的人群疏散机器人的运动规划算法。首先通过在原始的社会力模型中加入人机作用力构建出人机社会力模型，从而利用机器人对行人的作用力来影响人群的运动状态；然后基于DQN设计机器人运动规划算法，将原始行人运动状态的图像输入该网络并输出机器人的运动行为，在这个过程中将设计的奖励函数反馈给网络使机器人能够在"环境-行为-奖励"的闭环过程中自主学习；最后经过多次迭代，机器人能够学习在不同初始位置下的最优运动策略，最大限度地提高总疏散人数。在构建的仿真环境里对算法进行训练和评估。实验结果表明，与无机器人的人群疏散算法相比，基于DQN的人群疏散机器人运动规划算法使机器人在三种不同初始位置下将人群疏散效率分别增加了16.41%、10.69%和21.76%，说明该算法能够明显提高单位时间内人群疏散的数量，具有灵活性和有效性。     

Powered two-wheelers road accidents and their risk perception in dense urban areas: Case of Paris

For a few years, the use of powered two-wheelers has taken off in Paris. It then became critical for the City of Paris to understand both the mechanisms leading to traffic accidents involving at least one powered two-wheelers user and the perception of their risk when riding in dense urban areas. In so doing, two studies were carried out along similar lines so that their results could be compared. The first study focused on the perception of situations where accidents are most likely to occur. The second one was an analysis of police reports of accidents involving at least one powered two-wheelers and the drawing-up of prototypical accident scenarios. Comparing the results of the two studies revealed a gap between perceived and objective risks of these users. In fact, they rather fear the situations during which a car driver is changing lanes, while accidents involving them occur more often when a car driver turns (right, left or U). Knowledge of this dissonance in terms of awareness of road risks for powered two-wheelers and equally, other road users, will give the City of Paris food for thought. The promising results of this study have encouraged the City of Paris to extend it to other types of users, such as cyclists or elderly pedestrians.     

移动群智感知中融合数据的隐私保护方法

随着移动智能设备的普及,群智感知得到广泛应用,也面临严重的隐私泄露问题.现有隐私保护方案一般假设第三方服务平台是可信的,而这种假设对应用场景要求较高.基于此,提出了群智感知中一种新的数据融合隐私保护算法ECPPDA(privacy preservation data aggregation algorithm based on elliptic curve cryptography).服务器将参与者随机划分成g个簇,并形成簇公钥.簇内节点通过簇公钥加密数据并融合得到簇融合结果数据.服务器通过与簇内成员协同合作得到融合结果原文,由于服务器接收到的是融合密文且密文解密需要簇内所有节点共同协作,因此服务器不能得到单个参与者的数据.此外,通过服务器对簇公钥的更新,能够方便参与者动态加入或失效.实验结果显示ECPPDA具有高安全性、低消耗、低通信、高精度的特点.     

人群计数研究综述

人群计数广泛应用在公共安防、视频监控和智慧城市建设等领域,对控制特定场所人数、指挥公共交通、防止疫情蔓延、保障社会稳定具有重要积极意义。传统的计数方法精度不高、场景受限,随着深度学习的发展,传统方法逐渐被卷积神经网络（convolutional neural network,CNN）方法代替。介绍了人群计数的研究背景、现状和发展趋势,叙述了两种传统方法;从计数精度、网络结构、评价指标和数据集等方面重点分析了CNN方法,发现CNN技术可以有效解决多尺度和跨场景等问题;阐述了基于Vision Transformer（ViT）序列的弱监督计数方法并且对比各类方法。对未来人群计数的研究前景做出展望。     

稠密网格上交互指定局部区域的新方法

在造型、绘制以及动画中,在网格上指定局部区域都是一个重要的操作。传统的3D动画软件采用的交互方法在选取稠密网格上的特征局部区域时需要繁琐的用户操作,而现有的其他交互方法又难以做到实时响应。该文提出了一个在网格上指定局部区域的新方法,该方法基于网格上的近似最短路径算法。该算法不对网格进行全局加细,而是利用预处理阶段求出的信息对网格进行局部加细,提高了算法的效率。而且通过网格上近似等距线的抽取,还能指定网格上的拓扑同构于环形的窄带区域。     

数据驱动的人群场景动画合成方法

为了向虚拟城区环境加入复杂而灵活的人群动画,提出一种数据驱动的人群场景动画合成方法.通过采集各种典型场景下人群流动的视频,提取出agent状态响应数据集;并将这些状态响应数据集嵌入到虚拟城区知识模型中.最后基于该虚拟城区知识模型实现了一个虚拟城区中人群流动场景的动画合成系统.