7月26日，西湖大学智能无人系统实验室举办了一场题为“Composite Learning Tracking and Interaction Control for Compliant Robots”的讲座，主讲人为中山大学潘永平教授。潘教授凭借其丰富的学术经历和研究经验，深入浅出地介绍了该领域的前沿技术。
潘教授在讲座中重点介绍了由柔性执行器驱动的机器人在追踪和交互控制方面的主要研究成果。随着全球人口迅速老龄化的趋势，当前机器人研究正由传统的工业机器人转向以人为中心的机器人，这些机器人能够与人类共存、合作或协作，包括服务机器人。由于存在人机物理交互，我们通常为这些以人为中心的机器人引入柔顺性。潘教授团队提出了一种启发自小脑学习和控制机制的数据驱动在线学习方法，称为复合学习。该方法在一致学习和强鲁棒性方面取得了严格的理论结果，彻底改变了现有难以在线学习和易受攻击的自适应系统。随后，他们成功解决了复合学习在机器人应用中的一系列关键理论挑战，并将其成功应用于由柔性执行器驱动的机器人的轨迹跟踪和交互控制中，从而显著提升了这些机器人的精度、安全性和自然性。
图 讲座现场师生
本次学术交流不仅展示了复合学习在柔性机器人领域的创新应用，还深刻阐述了面向未来的机器人技术发展方向。通过潘教授的深入讲解和研究成果的分享，我们深入了解了在人口老龄化背景下，柔性执行器驱动的机器人如何通过先进的复合学习方法实现更精确、更安全、更自然的交互。这不仅对推动智能无人系统领域的科研进展具有重要意义，也为未来智能技术在社会福祉和产业发展中的广泛应用提供了新的思路和解决方案。我们由衷的感谢潘教授的到访，并期待未来有更多的学术交流与合作机会。
讲座预告信息如下
时间：2024年7月26日（周五）14:00-16:00
Time: 14:00-16:00, Friday, July 26, 2024
地点：云谷校区E10-304
Venue: E10-304, Yungu Campus
主持人: 工学院 赵世钰 博士
Host: Dr. Shiyu Zhao, School of Engineering
语言：中文
Language: Chinese
主讲人简介/Biography:
Yongping Pan is a Professor who leads the Intelligent Robotics Lab at the Sun Yat-sen University, Shenzhen, China. He holds a Ph.D. degree in control theory and control engineering from the South China University of Technology, Guangzhou, China, and has over ten years of research experience in top universities in Singapore and Japan. His research interests lie in automatic control and machine learning for robotics. He has authored or co-authored more than 180 peer-reviewed academic papers, with over 130 papers in refereed journals. His publications have attracted over 7400 and 5700 citations in the Google Scholar and Web of Science Core Collection, respectively. Dr. Pan is currently serving as the Chair of the IEEE Robotics and Automation Society Guangzhou Chapter and an Associate Editor of six top-tier journals published by IEEE and IFAC. He has served as an Organizing Committee Member of five international conferences and the Lead Workshop Organizer of the IEEE Conference on Decision and Control. He has been recognized as a Global Highly Cited Researcher by Clarivate, a Most Cited Chinese Researcher by Elsevier, and a World Top 2% Scientist (both single year and career) by Stanford University. Furthermore, he has been invited to deliver academic talks at leading universities and conferences over 60 times worldwide.
讲座摘要/Abstract:
With the rapid population aging globally, the current trend of robotic research has been shifting from traditional industrial robots to human-centered robots that coexist, cooperate, or collaborate with humans, including service robots. Due to the existence of physical human-robot interaction, we usually introduce compliance to human-centered robots. This talk introduces our major results in tracking and interaction control for robots driven by compliant actuators. First, we establish a data-driven online learning methodology termed composte learning inspired by the cerebellum learning and control mechanism, and its rigorous theoretical results on consistent learning and strong robustness, which revolutionizes existing adaptive systems that are difficult to learn online and vulnerable. Then, we solve a series of key theoretical challenges in the robotic applications of composite learning, and apply it to trajectory tracking and interaction control of robots driven by compliant actuators, which improves their overall accuracy, safety, and naturalness.
西湖大学官网讲座信息链接：Composite Learning Tracking and Interaction Control for Compliant Robots
