报告题目：可穿戴设备的人体运动动力学建模及基于机器学习的模型参数识别

报告人：梅旭涛研究员

报告时间：2022年11月2日19:30-21:30

报告地点：#腾讯会议：160-662-420 会议密码：2022

主持人：孙亚辉

报告摘要：This topic will present how to build the dynamic model of arm motion during human walking based on the dynamics. This talk firstly will give an introduction to arm motion during human walking, then based on the space dynamics derive a mathematical model of the arm swing motion. In addition, we will discuss the effects of the road slope, walking speed, human height and arm length on the acceleration of the human wrist. Based on the derived mathematical model, we will introduce the use of machine learning to obtain the parameters of the proposed mathematical model. Results demonstrate that the Random Forest regression (RFR) method has better performance for arm swing predictions. These findings provide theoretical guidance for further design and optimization of wrist-worn energy harvesters in the future.

专家简介：Dr. Xutao Mei is currently a project researcher at Institute of Industrial Science (IIS) of the University of Tokyo (UT), working on energy harvesting, nonlinear dynamic, data modeling and machine learning. Before that, he obtained his Ph.D degree from UT in Mar. 2020. He was a recipient of the MEXT Scholarship of Japanese Government from Oct. 2016 to Mar. 2020. In the scope of Scientific Research, Dr. Xutao MEI published 13 journal papers, including 9 first-author journal papers on the top journal such as Joule, Nonlinear dynamic, Mechanical Systems and Signal Processing and Journal of Sound and Vibration, and 8 papers in International Conference, covering energy harvesting, nonlinear dynamic and data modeling. In the presentation, Dr. Xutao MEI will introduce one research works on “Machine learning based arm motion model for wearable energy harvester”.