Mishik A. Kazaryan (1948–2020) – One of the pioneers of Russian hydrogen energetics

In 2018, Mishik Airazatovich Kazaryan received the highest award of the International Association for Alternative Energy and Ecology - Order of Antoine de Saint-Exupéry “For Improving the Quality of Life on the Planet of People” (IAAEE) on nominating the Award Committee of the Editorial Board of the International Scientific Journal for Alternative Energy and Ecology (ISJAEE). The award was given for his outstanding contribution to development of alternative energetics and ecology. M.A. Kazaryan's prominent contribution to the development of alternative energetics and ecology is based on his pioneering works in the field of development of methods for producing hydrogen as environmentally friendly safe fuel, as well as works in the field of processing organic compounds by various physical methods. As a part of joint research with colleagues from Lebedev Physical Institute of RAS (LPI), M.A. Kazaryan participated in creation of new methods for producing hydrogen from various chemical compounds. The method of conversion of liquid-phase compounds in plasma discharges under the influence of intensive ultrasonic cavitation occupies a special place. In the course of these works, it is shown that low-temperature plasma initiated in liquid-phase media in discharge between electrodes is able to effectively decompose hydrogen-containing molecules of organic compounds and form gaseous products where the part of hydrogen is more than 90%. Estimations of energy efficiency calculated taking into account hydrogen combustion heat and initial substances, as well as electricity costs, showed an efficiency level of about 60–70% in depending on the composition of the starting mixture. Another notable contribution of M.A. Kazaryan to the development of alternative energetics was the work on the optimization and justification of technological and structural parameters of energy discharge devices based on high-voltage pulse-periodic discharge for creating a reactor for plasmachemical processing of polymer wastes into hydrogen and other valuable compounds.     

Optimal filtering and control for wireless networked closed-loop control systems

This article studies the optimal filtering and control for wireless networked control systems (WNCSs). In WNCSs, packets may be lost in both control and feedback channels and user datagram protocol is usually used to improve the performance of the real-time control. Relevant literature indicates that the conventional optimal filtering for such a system cannot be applied in practice due to the complex calculation with Gaussian mixtures. This paper proposes a novel scheme to realize the optimal filtering and the linear quadratic Gaussian control for WNCSs, in which the controlled node performs a local estimation and the remote-control node performs the final estimation and control, and a synchronization of two estimators is guaranteed by a communication mechanism. An optimal filtering algorithm is developed, the stability condition of the filtering error covariance is obtained, optimal finite-horizon and infinite-horizon control are derived, and the stability of the closed-loop control system is proved. Numerical simulations show the validity and feasibility of the theoretical results.     

Creating formative feedback spaces in large lectures

Large lectures are the predominant way of teaching first-year students at universities in Norway. However, this forum for education is seldom discussed as a context for a formative feedback practice. The purpose of this sequential mixed methods study was to address whether and how a student-response system can open for a formative feedback practice in lectures and thereby support students' ability to monitor their own learning, as well as supply insight into how students engage with the feedback in their course work. The context for the study was large lectures (150–200 students) in a qualitative method course for first-year psychology students. Findings from the survey (n = 149) showed a positive correlation between the extent to which students report that they use clickers to monitor their own learning, and the extent to which they report that they used the feedback in their own course work. However, findings indicate that students valued the process of monitoring their own learning during the lectures to a greater extent than they actually used the feedback in their course work. Findings from interviews (n = 6) illustrated various ways students applied feedback in their course work.     

Classification and regression models of audio and vibration signals for machine state monitoring in precision machining systems

We present a data-driven method for monitoring machine status in manufacturing processes. Audio and vibration data from precision machining are used for inference in two operating scenarios: (a) variable machine health states (anomaly detection); and (b) settings of machine operation (state estimation). Audio and vibration signals are first processed through Fast Fourier Transform and Principal Component Analysis to extract transformed and informative features. These features are then used in the training of classification and regression models for machine state monitoring. Specifically, three classifiers (K-nearest neighbors, convolutional neural networks and support vector machines) and two regressors (support vector regression and neural network regression) were explored, in terms of their accuracy in machine state prediction. It is shown that the audio and vibration signals are sufficiently rich in information about the machine that 100% state classification accuracy could be accomplished. Data fusion was also explored, showing overall superior accuracy of data-driven regression models.     

基于深度学习的人体姿态估计方法综述

基于深度学习的人体姿态估计方法旨在通过构建合适的神经网络，直接从二维的图像特征中回归出人体姿态信息。主要按照2D人体姿态估计到3D人体姿态估计的顺序，并从单人检测与多人检测、稀疏的关节点检测与密集的模型构建等方面，对近年来基于深度学习的人体姿态估计方法进行系统介绍，从而初步了解如何通过深度学习的方法得到人体姿态的各个要素，包括肢体部件的相对朝向和比例尺度、骨骼关节点的位置坐标和连接关系，甚至更为复杂的人体蒙皮模型信息。最后,对当前研究面临的挑战以及未来的热点动向进行概述，清晰地呈现出该领域的发展脉络。     

Distributed learning consensus control based on neural networks for heterogeneous nonlinear multiagent systems

This paper considers a novel distributed iterative learning consensus control algorithm based on neural networks for the control of heterogeneous nonlinear multiagent systems. The system's unknown nonlinear function is approximated by suitable neural networks; the approximation error is countered by a robust term in the control. Two types of control algorithms, both of which utilize distributed learning laws, are provided to achieve consensus. In the provided control algorithms, the desired reference is considered to be an unknown factor and then estimated using the associated learning laws. The consensus convergence is proven by the composite energy function method. A numerical simulation is ultimately presented to demonstrate the efficacy of the proposed control schemes.     

深度学习识别光网络单元故障的设计与应用

为解决依赖装维上门鉴别光网络单元故障带来的不便,可以从机器视觉入手实现自动化故障识别。近年,ImageNet挑战赛的成功推动了物体识别技术的跨越式发展,特别是基于卷积的深度学习技术在视觉识别方面已经达到人类水平,为光网络单元故障的自动识别提供了技术基础。文章对识别光网络单元的工作状态进行了研究,将设备工作状态分为7个场景,提出了利用手机APP采集图片识别故障的解决方案并投入了实际生产;重点阐述了深度学习模块的设计与实现,提出一种通过算法整合的方式综合运用物体检测和图像分类算法,分3阶段逐步求精,解决了图片过滤,光网络单元型号和状态识别等问题,实现了基于计算机视觉自动识别光网络单元故障。从数据上看产品的端到端准确率超过84%,识别速度达到10 FPS,月均提供服务超过1万人次,在减少用户等待的同时节约了人力资源。     

基于深度学习的视频跟踪研究进展综述

近年来深度学习迅猛发展，颠覆了语音识别、图像分类、文本理解等领域的算法设计思路。深度学习因其具备强大的特征提取能力，在图像识别领域的成绩尤为突出。然而深度学习与视频监控领域的结合并不多，由于深度模型具有多层网络结构，算法复杂度大，训练和更新模型时比较耗时，很难满足实时性要求。回顾了深度学习的发展史，介绍了最近10年来国内外深度学习主要模型，论述了基于深度学习的目标跟踪算法，指出了各算法的优缺点，最后对当前该领域存在的问题和发展前景进行了总结和展望。     

叠合盆地凸起区多期复杂断裂特征及形成机制——以准噶尔盆地车排子凸起为例

以高精度三维地震资料为基础,结合构造演化恢复和区域构造背景,系统研究了准噶尔盆地车排子凸起断裂分布和形成机制。结果表明,车排子凸起发育深部和浅部2套断裂体系,深部断裂体系和浅部断裂体系分别发育3种断裂样式。车排子凸起共经历5个构造演化阶段,其中,晚二叠世挤压-逆冲控制了深部断裂体系的初始形成,晚侏罗世逆冲-走滑控制了深部断裂体系的最终定型,新近纪叠加掀斜、局部伸展控制了浅部断裂体系的形成。不同断裂体系在油气成藏过程中均起到了重要作用。其中:深部断裂体系的红车断裂是重要的油源断裂,低序级断层控制形成了石炭系火山岩优质储集体;浅部断裂体系使得油气向浅部层系运移调整,同时形成了断块圈闭、断鼻圈闭和断层-岩性圈闭等丰富的圈闭类型。     

Image and Feature Space Based Domain Adaptation for Vehicle Detection

The application of deep learning in the field of object detection has experienced much progress. However, due to the domain shift problem, applying an off-the-shelf detector to another domain leads to a significant performance drop. A large number of ground truth labels are required when using another domain to train models, demanding a large amount of human and financial resources. In order to avoid excessive resource requirements and performance drop caused by domain shift, this paper proposes a new domain adaptive approach to cross-domain vehicle detection. Our approach improves the cross-domain vehicle detection model from image space and feature space. We employ objectives of the generative adversarial network and cycle consistency loss for image style transfer in image space. For feature space, we align feature distributions between the source domain and the target domain to improve the detection accuracy. Experiments are carried out using the method with two different datasets, proving that this technique effectively improves the accuracy of vehicle detection in the target domain.