Andreas Daniel, Werner Dominik, Hou Zhongshi, Dwivedi Anany, Castellini Claudio, Beckerle Philipp
Chair of Autonomous Systems and Mechatronics, Friedrich-Alexander-Universität Erlangen-Nürnberg, 91054, Erlangen, Germany.
Artificial Intelligence (AI) Institute, Division of Health, Engineering, Computing and Science, University of Waikato, Hamilton, 3216, New Zealand.
Sci Data. 2025 Aug 28;12(1):1507. doi: 10.1038/s41597-025-05852-6.
This paper introduces MyoKi, a database capturing multimodal myography and hand kinematics during various realistic daily life activities. MyoKi emphasizes the complexity of real-world settings, addressing limitations of existing databases, which often reflect controlled laboratory conditions. The database includes two subsets of participants designed to evaluate different sensor configurations. Both subsets contain surface electromyography (sEMG) and inertial measurement unit (IMU) data, along with hand kinematics covering 18 finger and wrist joints. For the second subset, additional force myography (FMG) data was collected. The database captures hand movements of 35 participants performing 74 tasks, with varying arm orientations and movements involving different grips and motions. By offering detailed participant profiles and systematically categorizing each task, the MyoKi database enables in-depth exploration of task complexity, sensor influence, and the impact of demographic and anthropometric factors on control system performance. The database is designed to facilitate research in continuous hand control, enhancing the robustness and reliability of myoelectric devices for daily activities, moving towards user-friendly and effective control of robotic and prosthetic hands.
本文介绍了MyoKi，这是一个数据库，用于捕捉各种现实生活活动中的多模态肌电图和手部运动学数据。MyoKi强调现实世界环境的复杂性，解决了现有数据库的局限性，现有数据库往往反映的是受控的实验室条件。该数据库包括两个参与者子集，旨在评估不同的传感器配置。两个子集都包含表面肌电图（sEMG）和惯性测量单元（IMU）数据，以及涵盖18个手指和腕关节的手部运动学数据。对于第二个子集，还收集了力肌电图（FMG）数据。该数据库记录了35名参与者执行74项任务时的手部运动，包括不同的手臂方向以及涉及不同抓握和动作的运动。通过提供详细的参与者概况并对每个任务进行系统分类，MyoKi数据库能够深入探索任务复杂性、传感器影响以及人口统计学和人体测量学因素对控制系统性能的影响。该数据库旨在促进连续手部控制方面的研究，提高用于日常活动的肌电设备的鲁棒性和可靠性，朝着对机器人手和假肢手进行用户友好且有效的控制迈进。
