基于磁-惯性传感融合的灵巧手指单元全姿态信息反演方法

Method of Dexterous Finger Unit Full-pose Modeling Based on Magnetic-inertial Sensor Fusion

作为人类与外界环境互动的重要自然媒介,手具有高度的灵活性和复杂性,高效准确的手部动作和姿态识别对实现基于手部的人-机-环境共融具有重要意义。针对手部显著的个体差异特性,提出基于磁-惯性传感信息融合技术的个性化手指单元参数化模型构建方法,为实现精准的人-机-环境感知和共融提供技术支持。结合手部的指骨生物构造约束,建立手指单元运动学模型;通过结构化磁场标记技术,搭建磁-惯性信息融合的手指单元全姿态反演系统,完成个性化手指单元参数化模型构建。通过对不同志愿者手指单元开展连续路径对比试验以及单点定位精度验证试验,证明提出的基于磁-惯性传感信息融合技术的个性化手指单元全姿态信息反演方法定位精度分布在[1.43,2.81]mm。

As an important natural medium for human-environment-interaction,human hand takes advantages of high degrees of flexibility and complexity,and efficient and accurate hand motion and gesture recognition is of great significance for realizing hand-based human-machine-environment integration.A method for constructing a personalized dexterous finger unit full-pose information based on the magnetic-inertial sensor information fusion technology is proposed,in order to provide technical support for accurate human-machine-environment perception and inclusiveness.Combined with the biological structural constraints of the phalanx of the hand,a finger unit kinematics model was established.Through the structured magnetic field marking technology,a magnetic-inertial information fusion finger unit full-pose deriving system is developed,and the construction of a personal parameterized finger unit model is completed.The continuous path comparison test and single-point positioning accuracy verification test of different volunteer finger units proved that the proposed method for personalized finger unit full-pose information derivation based on magnetic-inertial sensor information fusion technology can achieve a positioning accuracy between[1.43,2.81]mm.