The team of the ISSDP laboratory participated in workshop “Smart Algorithms in Production” organized by SPbPU and the University of Aachen
On March 11, the second joint workshop of the two universities for young scientists on “Smart Algorithms in Production” took place.
The event was a logical continuation of the workshop on smart robotics in manufacturing, which was held in June 2020.
Innovative technologies, based on unique smart algorithms of data processing, are becoming more and more important in everyday life and solve problems in a wide variety of industries every year. The main goal of the “Smart Algorithms in Production” workshop was to establish new research contacts between young scientists and identify the potential for joint projects and research.
The team of SPbPU included: the Acting Director of the Institute of Computer Science and Technology, Professor Lev Utkin, researchers and graduate students of the Higher School of Cyber-Physical Systems and Control, the Higher School of Applied Mathematics and Computational Physics, the Higher School of Biomedical Systems and Technologies, the Industrial Systems for Streaming Data Processing Laboratory and the Laboratory for Modeling of Technological Processes and Design of Power Equipment of the SPbPU NTI Center, and representatives of international services.
Aachen University was represented by Dr. Matthias Brockmann, Head of the Advanced Internet Manufacturing Cluster, researchers of the Department of Control and Automation Technologies, Department of Information Systems, Institute of Textile Engineering, Institute of Information Management in Mechanical Engineering, representatives of the Process and Data Science Group.
Dr. Matthias Brockmann opened the seminar and told about the work of the advanced Internet manufacturing cluster, about the transfer of the Internet of Things into the world of manufacturing, and about the joint work between SPbPU and Aachen University for the past year, which, despite the epidemiological situation and the inability to meet in person, is actively developing.
On behalf of SPbPU, the participants were welcomed by the acting director of the Institute Professor Lev Utkin, who presented the main directions of the Institute’s activities and developments in the fields of artificial intelligence, machine learning, cyber-physical systems, etc., and gave an overview of the current projects of the Research Laboratory for Neural Networking Technologies and Artificial Intelligence.
The working part of the seminar was held in a pitch-session format. Researchers and young scientists of Polytechnic University and Aachen University made presentations of their scientific research and projects in a wide range of topics, united by the common theme of intelligent algorithms. In particular, the following points were discussed: artificial intelligence methods for analyzing an unstructured point cloud, visualization technologies, and deep learning for medical visualization, as well as the topics of information modeling in a real estate management system and a hybrid neural network model for industrial facility management and condition monitoring.
The staff of the ISSDP laboratory presented three reports. Dmitry Barinov, leading programmer of the Industrial Systems for Streaming Data Processing Laboratory of the SPbPU NTI Center, in his report “Possibilities of using a miniature fluorescence microscope” described his contribution to the project “Study of neural network activity changes, caused by neurodegenerative diseases, using methods of artificial intelligence,” implemented jointly with the Graduate School of Biomedical Systems and Technologies (GSBSiT) of the IBSiB SPbPU.
The technology of neuronal activity visualization in laboratory rodent brains in vivo using a miniscope integrated at the Laboratory of Molecular Neurodegeneration of the WSBSiT ISBSiB SPbPU required a new digital tool for mathematical analysis of the data obtained from the experiment. The software package NeuroInfoViewer, created in the ISSDP laboratory, allows for high-level analysis and construction of correlation dependence of neuronal activity.
Laboratory engineer Ivan Orlov in co-authorship with lead engineer Andrey Antonov made a report on “Precise clock synchronization project for industrial applications: a roadmap”, in which he continued the story about the way to achieve high precision synchronization of distributed systems objects. At the previous workshop “Smart Robotics in Production,” Ivan presented the developments of the laboratory for the use of the White Rabbit clock synchronization solution in the radio-frequency positioning of objects indoors. At the current workshop, he highlighted the further progress of the project: the results achieved on experimental samples, the process of creating several industrial solutions, and plans for further development of equipment with sub-nanosecond precision synchronization.
Nikolay Abramov, a leading programmer, an expert in training neural networks of the ISSDP Laboratory, in his report “Analysis of approaches to the detection of woven fabric defects” told about the development of methods and hardware-software tools to detect fabric defects in the textile industry using neural network methods (together with IVGPU). The project involves the development of methods and models of multitasking training of neural networks, as well as their software implementation for complex control. For this purpose, a classification of defects by prevalence and degree of influence on the grade of products will be made. At the moment, as part of Phase 1 of the project, an analysis of approaches to identifying defects in woven fabrics is being carried out.
“The possibility of interaction with the University of Aachen is very important. The past seminars and meetings showed a great variety of directions for international scientific cooperation and the high activity of young researchers in both Russia and Germany. I see here great prospects for the development of projects in many industries, including the light industry, construction, and medicine. Also, in my opinion, it will be very useful for young researchers to discuss their technical solutions with colleagues from Aachen, it will allow them to take a fresh look at their developments and improve them,” said Marina Bolsunovskaya, head of the Industrial Systems for Streaming Data Processing Laboratory at the SPbPU NTI Center.
RWTH Aachen scientists in their speeches shared their experience and developments in the field of dynamic modeling of industrial robots using neural networks with physical plausibility, discussed methods and components for semantic web services, possibilities of transfer learning for image-based quality control, self-optimizing machines for nonwoven production, promising technologies for mining in production, and other topics.
“Our next joint workshop between Polytechnic University and the Advanced Internet Manufacturing Cluster was a great success. We have a full year of sustained partnership under our belt. I would like to emphasize that interdisciplinary and international cooperation is especially important in the field of smart algorithms in manufacturing. SPbPU scientists and researchers are leaders in the field of smart algorithms internationally, and I very much appreciate and welcome our professional exchange,” emphasized Dr. Matthias Brockmann, head of Aachen University’s advanced Internet manufacturing cluster.
The participants showed mutual interest in each other’s research and projects and agreed to continue discussing a number of scientific spheres on the presented topics. At the end of the seminar, all participants will receive certificates prepared by Aachen University. Besides, the parties determined other joint activities for the nearest future: so, colleagues from Aachen University will take part in the international conference on cyber-physical systems and control, which is held by Polytechnic University together with Leibniz University of Hannover (Germany) and City, University of London (Great Britain) at the end of June.
Source: spbstu.ru
