The number of people with hand disabilities caused by stroke is increasing every year. Developing a low-cost and easy-to-use data glove to capture the human hand motion can be used to assess the patient’s hand ability in home environment. While a majority of existing hand motion capture methods are too complex to be used for patients in residential settings. This paper proposes a new sensor layout strategy using the inertial and magnetic measurement units and designs a multi-sensor Kalman data fusion algorithm. The sensor layout strategy is optimized according to the inverse kinematics and the developed hand model, and the number of sensors can be significantly reduced from 12 in conventional systems to 6 in our system with the hand motion being completely and accurately reconstructed. Hand motion capture experiments were conducted on a healthy subject using the developed data glove. The hand motion can be restored completely and the hand gesture can be recognized with an accuracy of 85%. The results of a continuous hand movement indicate an average error under 15% compared with the common glove with full sensors. This new set with optimized sensor layout is promising for lower-cost and residential medical applications.

中文翻译：

---

基于IMMU的新型数据手套，用于手部动作捕捉，传感器布局经过优化

每年由中风引起的手部残疾的人数在增加。开发低成本，易于使用的数据手套来捕获人的手部运动可以用来评估患者在家庭环境中的手部能力。尽管大多数现有的手部动作捕捉方法过于复杂，以至于不能用于居住环境中的患者。本文提出了一种使用惯性和磁测量单元的新型传感器布局策略，并设计了一种多传感器卡尔曼数据融合算法。传感器布局策略根据逆运动学和开发的手部模型进行了优化，并且传感器的数量可以从传统系统中的12个显着减少到我们系统中的6个，并且可以完全准确地重建手部运动。使用开发的数据手套在健康受试者上进行了手势捕获实验。可以完全恢复手势，并且可以以85％的精度识别手势。与具有完整传感器的普通手套相比，连续不断的手运动结果表明平均误差在15％以下。具有优化的传感器布局的这套新设备有望用于低成本和住宅医疗应用。