Young Scholars Forum
Assoc. Prof. Hongtian Chen, Shanghai Jiaotong University, China
Speech Title: Mamba for Time Series Prediction and Fault Detection
Abstract: Mamba is an efficient long-sequence model that has recently gained wide attention for its strong capabilities in long-term sequence analysis and prediction. However, conventional Mamba models face limitations when modeling high-order and coupled dynamics in real physical systems. This talk reviews the development of Mamba, analyzes its shortcomings in temporal prediction, and introduces a novel Mamba-based fault detection method with enhanced dynamic representation and adaptability. Experiments on the AG600 aircraft demonstrate robust performance across varying altitudes and environmental conditions, enabling accurate dynamic fault diagnosis.
Biography: Hongtian Chen received the B.S. and M.S. degrees in School of Electrical and Automation Engineering from Nanjing Normal University, China, in 2012 and 2015, respectively; and he received the Ph.D. degree in College of Automation Engineering from Nanjing University of Aeronautics and Astronautics, China, in 2019.
He had ever been a Visiting Scholar at the Institute for Automatic Control and Complex Systems, University of Duisburg-Essen, Germany, in 2018. Now he is a Post-Doctoral Fellow with the Department of Chemical and Materials Engineering, University of Alberta, Canada. His research interests include process monitoring and fault diagnosis, data mining and analytics, machine learning, and quantum computation; and their applications in high-speed trains, new energy systems, and industrial processes.
Dr. Chen was a recipient of the Grand Prize of Innovation Award of Ministry of Industry and Information Technology of the People's Republic of China in 2019, the Excellent Ph.D. Thesis Award of Jiangsu Province in 2020, and the Excellent Doctoral Dissertation Award from Chinese Association of Automation (CAA) in 2020. He currently serves as Associate Editors of a number of scholarly journals such as IEEE TII, IEEE TFS, IEEE TIM, CEP, etc.
Prof. Meng Zhang, Xi’an Jiaotong University, China
Speech Title: Understanding and Rethinking LiDAR-Inertial Odometry
Abstract: LiDAR-Inertial Odometry (LIO) has witnessed numerous novel datasets and approaches in recent years. Despite approaches’ progress and usefulness, few efforts have been devoted to thorough analysis regarding their non-determinism. Specifically, non-determinism refers to the property where a given input may produce different outputs, and the presence of unpredictable outcomes may distract and complicate the evaluation of a LIO algorithm's strengths and weaknesses. In this talk, we delve into the non-deterministic behavior of current state-of-the-art algorithms such as DLIO and FAST-LIO2, examining their design and implementation in detail. Then, solutions are provided to enable these algorithms to be deterministic. In addition, we propose improvements in submap building and the state update component of the LIO framework, resulting in an accurate and deterministic baseline: ADD-LIO. Extensive experiments have been performed on three challenging datasets, including SubT-MRS datasets, Newer College datasets and its extensions, covering a total trajectory length of approximately 17000 meters and span about 6 hours of data. The experimental results demonstrate that the proposed ADD-LIO consistently achieves high accuracy and determinism. To benefit the community, ADD-LIO and the deterministic version of DLIO and FAST-LIO2 are open-sourced.
Biography: Meng Zhang is a professor and doctoral supervisor at Xi'an Jiaotong University, a young Changjiang Scholar from the Ministry of Education. Meng Zhang obtained his Ph.D degree from the school of control science and engineering of zhejiang university. Meng Zhang has received honors such as the First Prize for Excellent Achievements in Science and Technology Research at Shaanxi Higher Education Institutions, the First Prize for Natural Science at the Chinese Association of Automation, the Outstanding Youth Award for Artificial Intelligence of Wu Wenjun. Meng Zhang has published more than 40 papers in journals such as Automatica, IEEE TAC Full Paper, IEEE TASE, etc. Meng Zhang serves as an associate editor of IEEE Transactions on Cybernetics, IEEE Transactions on Automation Science and Engineering and the chairman of IEEE IESONCON and other conference industrial forums. Meng Zhang’s research directions include intelligent control and optimization with applications to robotics and smart grids.
Assoc. Prof. Jingliang Sun, Beijing Institute of Technology, China
Speech Title: Cooperative Mission Planning Technique for Intelligent Swarms
Abstract: To address the development demands for autonomy, intelligence, and clustering of unmanned aerial vehicle (UAV) swarms, this report summarizes the current development status of swarm task planning at home and abroad. Then, the development context of collaborative task planning technologies is clarified. The main technical progress and engineering applications is introduced, such as UAV swarm task assignment, path planning, trajectory planning, and formation control. Finally, the key development directions of future collaborative task planning technologies for UAV swarms are prospected.
Biography: Jingliang Sun is an Associate Professor with the School of Aerospace Engineering, Beijing Institute of Technology. His research interests include multi-aircraft collaborative mission planning, decision-making, and cooperative guidance and control. He is in charge of 2 projects of National Natural Science Foundation of China and 4 National/provincial level projects and 8 other collaborative projects. He is selected in the Young Elite Scientists Sponsorship Program by BAST and the Beijing Nova Program. He has won one second prize of the Military Technology Invention Award and published 50 papers as the first author or corresponding author with 650 SCI citations except self-citations.
Prof. Xinghua Liu, Xi'an University of Technology, China
Speech Title: Deep Reinforcement Learning-based Service Restoration Strategy for Active Distribution Network
Abstract: The integration of distributed energy resources into distribution networks, marked by its inherent uncertainties, presents a substantial challenge for devising load restoration strategies. To tackle this challenge, we develop a memory-based graph reinforcement learning approach, designed to train the agent to acquire a critical load restoration strategy in a distribution network under uncertainties. Specifically, the restoration problem under uncertainties is formulated as a novel partially observable Markov decision process, where a multimodal feature-based observation space is proposed. This space includes graph-structured data of the environment and memory information of the agent. The graph-structured data contain potential features of the current observation, thus enhancing the observable domain, while the memory information incorporates temporal correlations between sample sequences to address the partial observability of the environment. Based on the proposed Markov process, we put forth a maximum entropy-based recurrent graph soft actor-critic algorithm to train the agent in partially observable environments through a recursive structure, where entropy regularization is utilized to facilitate a more extensive exploration of possibilities in a state space with high uncertainties.
Biography: Professor and PhD Supervisor at Xi'an University of Technology, recipient of the National Young Talent Program, Shaanxi Provincial Distinguished Young Scholar, Shaanxi Provincial High-Level Talent, Leader of the Shaanxi Provincial University Youth Innovation Team, and Head of the Shaanxi Provincial University Foreign Intelligence Base. He also serves as the Director of the Xi'an Key Laboratory of Cyber-Physical Power System Operation and Control. His main research focuses on networked control systems, state perception, security control, and resilience enhancement of new power systems, as well as optimal scheduling of integrated energy systems. To date, he has published over 100 SCI-indexed papers, including more than 50 articles in IEEE Transactions such as TAC, TCYB, TII, TSMC, TCASI, TITS, TTE, TASE, and TCASII, as well as top-tier energy journals. Among these, 10 papers are ESI Highly Cited Papers. He has authored two English monographs and holds 14 granted/applied national invention patents. He has led or co-led 24 national, provincial, and industry-sponsored projects, including collaborations with State Grid and other enterprises. His accolades include two provincial-level science and technology awards, 14 other academic and research awards, and over 10 teaching and educational honors.
Prof. Dongnian Jiang, Lanzhou University of Technology, China
Speech Title: Research on Key Technologies of Intelligent Perception and Optimization Decision in Nickel Smelting Process
Abstract: Based on the strategic needs of intelligent transformation and upgrading of the non-ferrous metal industry in Gansu Province, we focus on the core process of nickel pyrometallurgy. In response to the bottleneck problems of multi-source perception data loss, insufficient accuracy of process mechanism models, and lagging warning of key equipment failures in the smelting process of nickel flash furnaces and top blowing furnaces, we deeply integrate artificial intelligence and digital twin technology to carry out key technology research and industrial application of intelligent perception and optimization decision-making in the key process of nickel pyrometallurgy.
Biography: Dongnian Jiang, Professor and Doctoral Supervisor, awarded the "Longyuan Young Talents" talent project in Gansu Province and supported by the Outstanding Youth Fund of Gansu Province . He serves as a member of both the Technical Committee on Fault Diagnosis and Safety of Technical Processes and the Technical Committee on Predictive Control and Intelligent Decision. Dr. Jiang earned his undergraduate degree from Xiamen University (2006) and a Ph.D. in Control Theory and Control Engineering from Lanzhou University of Technology (2018). His research focuses on artificial intelligence, intelligent sensor design, and fault diagnosis. He has led two National Natural Science Foundation of China (NSFC) projects and contributed to over 20 major research initiatives including the National Key R&D Program, Gansu Outstanding Youth Fund Project, Provincial Key R&D Program, and Natural Science Foundation Key Projects. His scholarly output includes 40+ peer-reviewed publications in journals such as IEEE Transactions on Reliability, ISA Transactions, and Measurement.
Prof. Li Guo, Anhui Polytechnic University, China
Speech Title: Intelligent Fault Diagnosis for Heterogeneous Unmanned Swarm Systems
Abstract: The growing deployment of heterogeneous unmanned swarm systems in complex environments presents significant challenges in system reliability and maintenance. Key difficulties arise from system heterogeneity, dynamic operational conditions, communication uncertainties, and the intricate coupling between individual units and the swarm collective. These factors considerably complicate fault diagnosis and health management in practical applications. This talk will present our recently developed intelligent fault diagnosis frameworks that integrate knowledge-enhanced broad reinforcement learning to address these challenges. Finally, we will discuss future research directions in intelligent maintenance for autonomous swarm systems.
Biography: Guo Li received her Ph.D. in Communication and Information System from Sichuan University, Chengdu, China, in 2013. From 2016 to 2017, she has a visiting scholar at the University of Groningen in the Netherlands. She is currently a Professor and Vice Dean with the School of Electrical Engineering at Anhui Polytechnic University. She has been engaged in theoretical and applied research in data-driven fault diagnosis and health management, machine learning and image recognition for a long time, published over 60 academic papers in international academic journals and including 40 SCI journal articles, and granted 8 Chinese national invention patents. She is a member of the CAA and the CAAI , And she serves as a committee member of the Data-Driven Control, Learning and Optimization Committee and the System Intelligent Diagnosis and Health Management Committee.
Assoc. Prof. Yu Sun, Xi'an Jiaotong University, China
Speech Title: Multimodal Motion Soft Robotics for Damage Detection
Abstract: Multi-modal motion soft robots suitable for complex and narrow deep cavity environments can provide as an important carrier for intelligent inspection of narrow space and complex environment. Inspired by origami and bistable structures, this paper proposes a Kresling origami multi-motion modal crawling robot, a wall-climbing robot based on rigid-soft hybrid suction cups, and a miniature jumping robot based on biased bistability, introduces the corresponding structural design, gait control, and principle validation, and looks forward to the potential and optimization direction for future use in the inspection of the interior of aero-engine.
Biography: Yu Sun, an Associate Professor at the School of Mechanics, Xi'an Jiaotong University. In recent years, she has conducted research in the field of aero-engine blade damage mechanism and intelligent inspection robot. She has published more than 50 papers in journals such as Advanced Science, TRO, ASS, CPB, RAL, TIM, etc., among which 26 papers are SCI papers as first/corresponding authors; one academic publication related to multi-scale mechanics has been published; and 17 patents have been authorized. The research results have been tracked by ScienceNet, China Science and Technology Network, Phoenix New Media Science and Technology Channel and reprinted by Shaanxi Provincial Department of Industry and Information Technology, Chongqing University Institute of Science and Technology Development, and reported by Shaanxi News Broadcast.