基于GF-1数据的三峡库区水体信息精细化提取

为实现江河类狭长型水体信息的精细化提取，利用GF-1卫星数据，采用支持向量机和目视解译相结合的方法对三峡库区及重庆市水体信息进行了精细化提取。使用总体分类精度、Kappa系数、错分误差、漏分误差、制图精度和用户精度等指标对库区水体信息粗提取结果进行验证分析。结果表明:4个试验区水体提取的总体分类精度均超过90%，除试验区4的Kappa系数为0.884 1以外，其余试验区均超过0.9，提取精度较高。结合目视解译的方法，在粗提取结果的基础上对各问题进行精细化处理，得到精度高、完整性好的三峡库区以及重庆市水体信息数据，可为后续该区域的精细化遥感业务开展提供有效资料。     

Flight control of a hexa-rotor airship: Uncertainty quantification for a range of temperature and pressure conditions

The present paper is concerned with the dynamic modeling and design of control laws for a small non-rigid multi-rotor airship constituted of an oblate-spheroid helium balloon coupled with an electric-powered hexa-rotor airframe. The vehicle is assumed to operate in windless and low-speed conditions. A six-degree-of-freedom nonlinear dynamic model is derived for it using the Newton–Euler approach and considering, among other efforts, a restoring torque due to the displacement of the balloon’s center of buoyancy above the vehicle’s center of mass and the added-mass effect resulting from the air–structure interaction. Using the derived model and assuming a time-scale separation between the translational and rotational dynamics, the attitude and position control laws are designed separately from each other. Both laws are formulated using feedback linearization combined with control input saturation within appropriate parallelepipedal sets, which are carefully chosen to respect pre-defined bounds on the control torque, control force and maximum inclination angle. The effect of temperature and pressure fluctuations is taken into account through a parametric probabilistic approach, where Maximum Entropy Principle is used to construct a physically consistent stochastic model and Monte Carlo method is used as the stochastic solver to propagate the uncertainties through the system. Extensive simulation results show the effectiveness of the proposed control system and quantify the uncertainty of its performance over a wide range of local temperature and pressure.     

结合深度学习与相关反馈的遥感图像检索

目的 针对基于内容的图像检索存在低层视觉特征与用户对图像理解的高层语义不一致、图像检索的精度较低以及传统的分类方法准确度低等问题，提出一种基于卷积神经网络和相关反馈支持向量机的遥感图像检索方法。方法 通过对比度受限直方图均衡化算法对遥感图像进行预处理，限制遥感图像噪声的放大，采用自学习能力良好的卷积神经网络对遥感图像进行多层神经网络的监督学习提取丰富的图像特征，并将支持向量机作为基分类器，根据测试样本数据到分类超平面的距离进行排序得到检索结果，最后采用相关反馈策略对检索结果进行重新调整。结果 在UC Merced Land-Use遥感图像数据集上进行图像检索实验，在mAP（mean average precision）精度指标上，当检索返回图像数为100时，本文方法比LSH（locality sensitive Hashing）方法提高了29.4%，比DSH（density sensitive Hashing）方法提高了37.2%，比EMR（efficient manifold ranking）方法提高了68.8%，比未添加反馈和训练集筛选的SVM（support vector machine）方法提高了3.5%，对于平均检索速度，本文方法比对比方法中mAP精度最高的方法提高了4倍，针对复杂的遥感图像数据，本文方法的检索效果较其他方法表现出色。结论 本文提出了一种以距离评价标准为核心的反馈策略，以提高检索精度，并采用多距离结合的Top-k排序方法合理筛选训练集，以提高检索速度，本文方法可以广泛应用于人脸识别和目标跟踪等领域，对提升检索性能具有重要意义。     

Research on Circular Area Search algorithm of multi-robot service based on SOA cloud platform

Considering autonomous mobile robots with a variety of specific functions as a kind of service, when there are many types and quantities of services and the density of regional services is large, proposing an algorithm of Circular Area Search (CAS) because of the problem of multi-robot service scheduling in various areas. Firstly, Django is used as the web framework to build the Service-Oriented Architecture (SOA) multi-robot service cloud platform, which is the basic platform for multi-service combination. Then, the service type, the latitude and longitude and the scoring parameters of the service are selected as the service search metrics to design the CAS algorithm that based on the existing service information registered in MySQL and the Gaode Map for screening optimal service, and then providing the service applicant with the best service. Finally, the service applicant applies for the self-driving tour service as an example to perform performance simulation test on the proposed CAS algorithm. The results show that the CAS algorithm of the multi-robot service cloud platform proposed in this paper is practical compared to the global search. And compared with the Greedy Algorithm experiment, the service search time is reduced about 58% compared with the Greedy Algorithm, which verifies the efficiency of CAS algorithm.     

基于改进人工蜂群算法的多目标绿色柔性作业车间调度研究

针对多目标绿色柔性作业车间调度问题（MGFJSP）的特点，提出从碳排放量、噪声和废弃物这3个指标来综合评定环境污染程度，建立了以最小化最大完成时间和环境污染程度为优化目标的MGFJSP模型，并提出了一种改进的人工蜂群算法来求解该模型。算法的具体改进包括：设计了一种三维向量的编码和对应解码方案，在跟随蜂搜索阶段引入一种有效的动态邻域搜索操作来提高算法的局部搜索能力，在侦查蜂阶段提出产生新食物源的策略用于增加种群的多样性。最后进行了实验研究与算法对比，以验证所建模型和所提算法的有效性。     

破碎围岩回采巷道锚网索支护参数优化

针对镇城底煤矿8"煤层工作面巷道围岩破碎、支护困难等问题,分析了原有巷道支护方式失效的原因,进而针对性的优化了巷道支护参数,采用①20 mm X 2500 mm左旋螺纹钢锚杆,①17.8 mm X 6500 mm锚索,并进行了验算。最后通过现场实测巷道顶板离层量得到有效的控制,实现了安全髙效开采。     

A case study of socio-cultural and technical factors in automobile design: Discourses between designers and potential users on a new electric vehicle in Africa

Technological artifacts emerge and develop within social contexts, inextricably interconnecting technology and society. This article addresses this interconnection empirically, analyzing the implementation of an electric vehicle designed and built by students of automotive engineering at the Technical University of Munich (TUM) in Germany. The students created this vehicle to address problems of mobility for rural populations in Africa, aiming to provide local residents with better access to necessities of life such as healthcare, education, and transportation. The article examines African perceptions of and responses to the vehicle, as well as specific suggestions for future improvement. It discusses not only African reactions to the vehicle and its potential social impact on African societies, but also the concepts and ideas of the designers behind this automotive construction. It also acknowledges the ways in which the engineers based their work on assumptions about African needs rather than on substantial ethnographic research, which diminished their ability to succeed.     

模型驱动的大数据流水线框架PiFlow

复杂流程的大数据处理多依托于流水线系统，但大数据处理的流水线系统在易用性、功能复用性、扩展性以及处理性能等方面存在不足。针对上述问题，为提高大数据处理环境的构建与开发效率,优化处理流程，提出了一种模型驱动的大数据流水线框架PiFlow。首先，将大数据处理过程抽象为有向无环图；然后，开发一系列组件用于构建数据处理流水线，并设计了流水线任务执行机制。同时，为规范和简化流水线框架的描述，设计了基于模型驱动的大数据流水线描述语言——PiFlowDL，该语言以模块化、层次化的方式对大数据处理任务进行描述。PiFlow以所见即所得(WYSIWYG)的方式配置流水线，集成了状态监控、模板配置、组件集成等功能，与Apache NiFi相比有2~7倍的性能提升。     

Experience-Based Cognition for Driving Behavioral Fingerprint Extraction

Abstract

With the rapid progress of information technologies, cars have been made increasingly intelligent. This allows cars to act as cognitive agents, i.e., to acquire knowledge and understanding of the driving habits and behavioral characteristics of drivers (i.e., driving behavioral fingerprint) through experience. Such knowledge can be then reused to facilitate the interaction between a car and its driver, and to develop better and safer car controls. In this paper, we propose a novel approach to extract the driver’s driving behavioral fingerprints based on our conceptual framework Experience-Oriented Intelligent Things (EOIT). EOIT is a learning system that has the potential to enable Internet of Cognitive Things (IoCT) where knowledge can be extracted from experience, stored, evolved, shared, and reused aiming for cognition and thus intelligent functionality of things. By catching driving data, this approach helps cars to collect the driver’s pedal and steering operations and store them as experience; eventually, it uses obtained experience for the driver’s driving behavioral fingerprint extraction. The initial experimental implementation is presented in the paper to demonstrate our idea, and the test results show that it outperforms the Deep Learning approaches (i.e., deep fully connected neural networks and recurrent neural networks/Long Short-Term Memory networks).     

Numerical study of flow distribution uniformity for the optimization of gradient porosity configuration of porous copper fiber sintered felt for hydrogen production through methanol steam reforming micro-reactor

A macroscopic numerical method is proposed to study the flow distribution uniformity of a novel porous copper fiber sintered felt (PCFSF), which has gradient porosities and was developed as the methanol steam reforming micro-reactor catalyst support for hydrogen production for fuel cell applications. The macroscopic porous media developed by the ANSYS/FLUENT software is used to represent the PCFSF. Our results indicate that the gradient porosity can reshape the flow distribution of PCFSFs greatly, thus producing significant influence on their performance. It is further revealed that, for a PCFSF with a determined gradient porosity configuration but different reactant feed directions, the velocity uniformity can be used as a quantitative criterion to evaluate the performance of hydrogen production. Furthermore, new gradient PCFSFs are produced according to the flow distribution of original gradient PCFSFs. The preliminary experimental results of the new gradient PCFSFs of 0.8-0.9-0.7 and 0.7-0.9-0.8 exhibit better methanol conversion and H₂ flow rate. This indicates that the numerical method can be used for the optimization of PCFSFs' gradient porosity configuration, which consists of the shape and position of the interfaces between different porosity portions, the number of interfaces and the porosity distribution in different portions.