Prof. Juanjuan Zhang
Nankai University, China
Biography: Dr. Juanjuan Zhang is a professor of College of Artificial Intelligence, Institute of Rototics & Automatic Information System, and Academy for Advanced Interdisciplinary Studies at Nankai University, China. She received her BEng on Electrical and Electronic Engineering with first class honors from Nanyang Technological University in 2007, and her MEng & PhD on Mechanical Engineering from Carnegie Mellon University in 2016. She joined Nankai University in 2017. She is a leading researcher in the field of 'human-in-the-loop' optimization for human-robot interactive systems. Her pioneering work was the first to make lower-limb exoskeleton assistance effective across a broad spectrum of users (Zhang et al., Science 2017), addressing the longstanding challenge of individual variability. She has served as Principal Investigator on multiple projects funded by the National Natural Science Foundation of China and the State Key Development programme of China. Dr. Zhang's current interests include human-robot interaction, 'human-in-the-loop' optimization, exoskeletons, rehabilitation robots, surgical robots, biomechanics, control theory and AI.
Speech Title: Human-in-the-Loop Optimization of Exoskeletons and Beyond
Abstract: Leveraging human responses to optimize and personalize human–robot interaction effectively closes a high-level control loop, enabling adaptive and individualized assistance that significantly enhances the performance of coupled human–machine systems. This paradigm has emerged as a powerful approach to address the long-standing challenge of inter-individual variability, particularly in wearable robotics such as lower-limb exoskeletons. This talk reviews the origin and evolution of human-in-the-loop optimization, highlighting key methodological advances and experimental breakthroughs that have enabled its practical deployment. It further examines the generalization and expansion of this paradigm beyond exoskeletons to a broader class of human-centered robotic and AI systems, where humans are not merely users but integral components of the control and learning process. Current limitations are discussed, including challenges related to efficiency, robustness, and scalability in real-world applications. Finally, emerging research directions at the intersection of biomechanics, control theory, and artificial intelligence are outlined, providing a unifying perspective on how human-in-the-loop approaches can drive the next generation of adaptive, personalized, and intelligent human–machine systems.
Prof. Zhan Li
Southwest Jiaotong University, China
Biography: Zhan Li received the B.Sc. and M.Sc. degrees in automation and communication and information systems from Sun Yat-sen University, Guangzhou, China, in 2005 and 2009, respectively, and the Ph.D. degree in robotics from University of Montpellier and INRIA, Occitanie, France, in 2014. He is currently a Professor with the Institute of Smart City and Intelligent Transportation (ISICT), Southwest Jiaotong University, Chengdu, China. Before joining Southwest Jiaotong University, he was a Senior Lecturer in the Department of Computer Science at Swansea University, Wales, U.K. His research interests include kinematic control of robots, neural networks, computational intelligence, rehabilitation engineering. He is currently serving as Associate Editors on many international journals. He has published over 90 papers including IEEE TNNLS, TSMCA, TII and so on.
Speech Title: Robotic Time-Optimal Motion Planning and Its Neurodynamics-Driven Control
Abstract: Time-optimal motion planning is critical for maximizing robotic operational efficiency, yet it poses significant challenges due to heterogeneous limits and constraints. Traditional approaches often suffer from computational complexity or lack of real-time adaptability. This speech presents a unified framework that integrates time-optimal path parametrization with neurodynamics-driven control. An efficient neurodynamics-based control law is developed to ensure fast, smooth, and stable tracking of the planned trajectory. The neurodynamic controller features inherent noise tolerance, model simplicity, and asymptotic convergence, enabling real-time disturbance rejection and constraint satisfaction. Simulation and experimental results on robotic manipulators and mobile robots demonstrate that the proposed approach achieves near-minimum-time performance while maintaining smooth control signals and robust stability. This work offers a practical and efficient solution for high-speed robotic operations in manufacturing, autonomous driving, and service robotics.
Assoc. Prof. Bo Sheng
Shanghai University, China
Biography: Associate Professor Bo Sheng, a Shanghai Leading Talent (Overseas) and Shanghai Pujiang Talent, obtained his Ph.D. and completed postdoctoral research at the University of Auckland, New Zealand, where he studied under international leading scholars in rehabilitation robotics and biomechanics: Professor Shane Xie (Fellow of both the Royal Society of New Zealand and Engineering New Zealand, National High-Level Leading Talent), Associate Professor Yanxin Zhang (Director of the Biomechanics Laboratory at the University of Auckland, Distinguished Minjiang Scholar of Fujian Province), and Associate Professor Lihua Tang (President of the New Zealand Chinese Scientists Association). Currently serving as a Tenured Associate Professor and Master’s Supervisor at the School of Mechatronic Engineering and Automation, Shanghai University (a Double First-Class university), his research focuses on bio-mechatronics integration systems, with key directions including large language model-driven smart elderly care/healthcare, intelligent human-robot interaction, rehabilitation medical robots, multimodal information fusion, and intelligent recognition algorithms based on electromyography (EMG) and electroencephalography (EEG). Centering on the mechanisms of complex human motion, intelligent control strategies, and cross-modal biosignal interpretation, his work aims to advance AI-enabled rehabilitation engineering and smart medical devices. To date, he has presided over or participated in more than 30 national and provincial-level research projects, published over 60 high-impact papers, and accumulated more than 1,200 total citations.
Speech Title: Development of Screening Tools for Adolescent Idiopathic Scoliosis Based on Deep Visual Algorithms
Abstract: In recent years, the incidence of postural issues among adolescents has been rising annually, drawing increasing attention to postural health. Traditional posture assessment primarily relies on manual evaluation by physicians, which is subject to bias and lacks sufficient accuracy. With advances in computer vision and deep learning, machine vision-based posture assessment methods have emerged. This study develops an assessment system for comprehensive multi-perspective analysis of adolescent posture using deep learning techniques. The system employs an enhanced key point detection model and an image segmentation model to extract critical postural information from multi-perspective posture images. On a self-constructed dataset containing multi-perspective images of 220 adolescents, the key point detection model achieved prediction accuracies of 84.9%, 77.9%, 88.7% and 85.3% for back, side-view with arms down, side-view with crossed arms, and overhead images, respectively; the segmentation model attained a prediction mAP of 76.2% on back images. After obtaining postural information, this study proposes 14 posture metrics based on key points and segmentation data, along with 7 diagnostic outcomes. The posture assessment system analyzes the postural information using these metrics and provides comprehensive evaluation results through a fuzzy matrix. In conclusion, the proposed deep learning-driven approach offers an objective and effective tool for posture assessment, enhancing the accuracy of evaluations and the identification of postural issues.
Assoc. Prof. Jianping Dou
Southeast University, China
Biography: Jianping Dou joined Southeast University, Nanjing, China, in 2009, and is now an Associate Professor and Ph.D. supervisor in School of Mechanical Engineering. He has undertaken four projects funded by the NSFC, and finished six contracted projects. He has published 50+ journal papers appeared in IEEE TSMCS & ITS, IJPR, JIM etc., with H-index 17. His current research interests include intelligent manufacturing, and evolutionary algorithm and its application.
He received his B.S. degree in Aircraft Propulsion Engineering from Nanjing University of Aeronautics and Astronautics, Nanjing, China, in 2002, M.S. degree in Aeroengine Control Engineering from Nanjing University of Aeronautics and Astronautics, Nanjing, China, in 2005, and the Ph.D. degree in Control Theory and Control Engineering from Southeast University, Nanjing, China, in 2009. He had been a visiting scholar in Department of Mechanical Engineering, University of Michigan (Ann Arbor) from Sept. 2016 to Oct. 2017.
Speech Title: An adaptive genetic algorithm with parameter control for operation sequencing in CAPP
Abstract: In Industry 4.0, the high volatility of customized products results in Operation Sequencing (OS) as a central module of CAPP being subject to frequent demands. To effectively address practical OS problems, which are inherently NP-hard, an adaptive Genetic Algorithm (GA) with robust performance and without requiring tedious parameter tuning is essential. Nevertheless, that adaptive GA is absent until now. In this talk, we intoduce such an adaptive GA with parameter control integrated the feedback information of population fitness and dynamic Fitness Landscape Analysis (FLA). For the AGA, the direct encoding of FOS and precedence-constraint satisfaction method are adopted, so that the AGA searches in the feasible solution space. In the AGA, the mutation rate is online updated by the population diversity, and four crossover operators are adjusted in a greedy fashion via the judgment combining population diversity and evolvability. To improve computational efficiency and enhance exploitation ability, the path-relinking local search is adaptively applied by the predefined maximum unchanged generation of best fitness. The AGA is compared with state-of-the-art algorithms including particle swarm optimization, harmony search, two recent GAs and a self-adaptive GA to verify its effectiveness and merits against five OS cases and six randomly generated OS problems. The comparative results show that the AGA outperforms existing MAs for solution quality, and illustrates its robust global search ability. |
