Prof. Dr. Xiaofang Yuan, Hunan University
Speech Title：Collaborative Operation Optimization of Processing Equipment and AGVs for Green Discrete Manufacturing
Abstract:With the energy crisis and environmental problems becoming serious, green manufacturing has attracted growing attention from academic and industrial communities. The discrete manufacturing process often involves a variety of production resources including processing equipment, transportation equipment, etc., and is characterized by strong coupling, nonlinearity, and multi-objectives. Collaborative optimization of production resources is one of the most critical means to globally improve production indicators. Based on the analysis of the needs and opportunities of collaborative optimization in the context of green discrete manufacturing, this report takes the research on the optimization of the collaborative operation of processing equipment and AGV as an example to introduce the relevant optimization model construction and solution methods.
Yuan Xiaofang, PhD, Professor in Hunan University, doctoral supervisor. He is mainly engaged in intelligent control theory and application, electric vehicle control and other aspects of research work. He has presided over more than 10 projects, including general projects of National Natural Science Foundation of China, youth projects of National Natural Science Foundation of China, special postdoctoral projects of China, science and technology plan of Hunan Province and Doctoral Program Fund of Ministry of education.He has more than 50 papers have been published in IEEE Transactions and other international authoritative journals in the field of control, including 40 SCI papers.
Prof. Dr. Yigang Cen, Beijing Jiaotong University
Speech Title：Self-RoI: A Self-perception Large Language Model for 3D Region-of-Interest Captioning
Abstract: Multi-modal Large LanguageModels (MLLMs) have enhanced practicality in 3D scene comprehension for real world applications. They have showcased promising performances in scene-level vision comprehension by visual instruction tuning. In this work, we extend MLLMs to the 3D Region-of-Interest (RoI) captioning task by enhancing their region-level vision comprehension ability. However, the scarcity of 3D RoI captioning datasets could constrain the performance of MLLMs. Moreover, constructing extra instruction datasets is often a resource-intensive and costly process. To enhance the performance of MLLMs with limited data, this paper introduces a method named self perception LLM for 3D RoI captioning (Self-RoI), which generates captions by integrating 3D visual representations and the associated implicit textual information of RoI objects. The implicit textual information is extracted by LLMs according to the corresponding visual representation. Furthermore, we propose a two-stage training scheme for maximizing data utilization. Specifically, after the standard alignment stage, we introduce the 3D RoI vision-caption interaction stage, which incorporates a semantic consistency regularization mechanism between the implicit textual information and corresponding human captions, thus guaranteeing that the generated captions from Self-RoI accurately describe the same target from diverse aspects. Experimental results confirm that our Self-RoI effectively comprehends region-level point clouds and markedly enhances the 3D RoI captioning performance over previous methods. Our code will be made available for further research.
Yigang Cen is a Professor and Deputy Director of the Institute of Information Science at Beijing Jiaotong University. He specializes in Internet of Things engineering and has made significant contributions in his field. He has led numerous research projects, including a Key Project of International Scientific and Technological Innovation Cooperation Between China and Serbia, as well as various national, provincial, and ministerial projects such as the National Natural Science Fund. Prof. Cen has published nearly 100 academic papers and has received several prestigious awards for his work. These include the "Special Award of Zhantianyou Railway Science and Technology Award," the "Excellent Thesis Award from IEEE Computer Society in 2020 (Second Runner-up)," the "Watch Teaching Grant of Beijing Jiaotong University," the "Excellent Tutor Award from Beijing Image Graphics Society," and the "Technology Application Award of Jiangsu Artificial Intelligence Society in 2021." Recognized for his expertise and achievements, Prof. Cen has been selected as the Class I Talent for the "Young Talents Program of Beijing Jiaotong University," the "'Double Hundred' Talent Cultivation Program of Beijing Jiaotong University Red Orchard," and has been honored as an "Advanced Individual of Three-Education in Beijing Jiaotong University." His guidance has also led doctoral students to win the excellent doctoral thesis award from the Beijing Image and Graphics Society in 2018, and graduate students to achieve success in competitions such as the China Postgraduate Mathematical Modeling Competition, Shanghai BOT Computer Vision Competition, and the 2018 IEEE Big Data Cup. Yigang Cen's primary research interests lie in the fields of computer vision, pattern recognition, intelligent transportation, and intelligent security.
Prof. Dr. Wei Li, Shanghai Institute of Microsystems and Information Technology, Chinese Academy of Sciences
Speech Title：Key sensing technologies and applications for power grid intelligent sensing
Abstract: The digital technology of new power systems is developing rapidly, and high-performance, low-cost sensors based on large-scale manufacturing technologies will undoubtedly play a key role. Image sensors, temperature sensors, current/voltage sensors, infrared sensors, gas sensors, and local discharge sensors are used to monitor the status of power facilities and equipment. Integrated with superior intelligent fault diagnosis algorithms, they are conducive to timely detection of security risks and ensure the safety of power grid operation. Starting from the intelligent perception of distribution network, this report introduces the key problems and trends existing in the current stage of power grid operation monitoring, and reports our main solutions and application cases from the perception layer, transmission layer and application layer. Finally, the future of intelligent perception technology of power distribution is prospected according to the current international and domestic technology frontiers.
Li Wei, an expert in micro-nano devices, director of the Joint Laboratory of Intelligent Perception of Energy Internet, He has presided over a number of scientific research projects as a project leader, including national H863 special key projects, national key research and development program topics, etc., and achieved a series of results with international influence.
Prof. Dr. Zhenmin Zhu, East China JiaoTong University
Speech Title：Study on the technology of 3D precision measurement based on polarization-coding structured light
Abstract: The high precision calibration of visual imaging and sensor is the core basic technology and fundamental guarantee of optical precision measurement. With the application of optical measurement gradually developed from detection in a fixed environment to measurement in an uncontrollable environment (such as field measurement, etc.), the optical properties of the measurement environment and the surface of the object to be measured become more and more complex. How to ensure the accuracy of the measurement system in the complex measurement environment has become an important direction of the development of structured light measurement system. At present, most of the researches on 3D optical measurement use light intensity as information transmission channel. Aiming at the problem of high precision measurement in complex environment, a calibration method of structured light system based on optimal polarization angle is proposed to improve the precision of system calibration. A single image field calibration method for all parameters of structured light sensor based on cylindrical target is proposed, which solves the high precision field calibration of in-situ structured light optical measurement system and promotes the engineering application of structured light measurement. A structured light coding technique based on polarization state is proposed to reduce the influence of camera defocus in structured light measurement system. A fast measurement technique of polarization enhanced structured light fringes is proposed, which can reconstruct a wide range of reflectivity objects with only once exposure time.
Professor Zhenmin Zhu (doctoral supervisor, April, 1984), leader of the doctoral program of Control Science and Engineering. On the Traffic Video Special Committee, Visual Inspection Special Committee of China Graphics and Image Society committee, Standing Director of Jiangxi province Automation Society. He has been selected as the Funding Program for Outstanding Young Talents of Jiangxi Province, Young Jinggang Scholar, the visiting scholar of "Light of the West" of organization Department of the CENTRAL Committee of the COMMUNIST Party of China, the Hundredpeople Voyage Project, etc. Serve as the evaluation expert of science and technology projects of National Natural Science Foundation of China, and Jiangxi province, Shanxi Province and Anhui Province. Presided over the research or completion of 2 National Natural Science Foundation projects and 10 of Jiangxi Outstanding Young Talents Project, key project of Industrialization of Provincial Invention Patent, provincial Natural Science Fund and other provincial and ministerial level and enterprise entrusted projects. Now, he mainly engaged in the research of precision vision measurement, intelligent detection and automation system, LED applied optics, etc. He has published more than 50 research papers, which 28 indexed in SCI, on international authoritative journal of optical measurement as Optics Express, Optics and Lasers in Engineering as the first author or corresponding author. He has applied for 24 national invention patents as the first author. The project achievement "Key Technology and Application to Improve visual imaging contrast" won the second prize of Jiangxi Provincial Technical invention, and participated in the second prize of provincial graduate Teaching Achievement and provincial Scientific and Technological Achievement. He is now the deputy dean of the School of Electrical and Automation Engineering, East China JiaoTong University.
Copyright© ICEERT 2023
2023 3rd International Conference on Information Control, Electrical Engineering and Rail Transit http://www.iceert.org/