基于数据手套与虚拟模型的手臂动作识别方法  

Arm Motion Recognition Method Based on Data Gloves and Virtual Model

作　　者：曹尔凡 崔建伟[1] CAO Er-fan;CUI Jian-wei(School of Instrument Science and Engineering,Southeast University,Nanjing 210096,China)

机构地区：[1]东南大学仪器科学与工程学院,江苏南京210096

出　　处：《测控技术》2021年第1期16-21,77,共7页Measurement & Control Technology

基　　金：国家自然科学基金项目(61873063)。

摘　　要：为了配合残疾人使用助残假手完成日常活动,提出了一种基于数据手套与虚拟模型的手臂动作识别方法。通过数据手套上的MEMS传感器测量手臂运动数据,结合虚拟交互环境中的手臂模型,完成模型对真实手臂的动作复现,并获取手臂模型的位置数据。而后将传感器加速度、角速度等数据的时域特征与手臂模型位置数据的动作轨迹特征结合起来,使用支持向量机算法对预定义的5种手臂动作进行分类识别,识别率达到99.33%,验证了提出的手臂动作识别方法的可行性。In order to cooperate with disabled people to use hand-assisted prosthetic hands to complete daily activities,a method of arm motion recognition based on data gloves and virtual models is proposed.The arm motion data was measured by the MEMS sensors on the data glove,combined with the arm model in the virtual interactive environment,the motion of the model to the real arm was completed,and the position data of the arm model was obtained.Then,the time domain characteristics of sensor acceleration,angular velocity and other data were combined with the motion trajectory characteristics of the arm model position data,and the support vector machine(SVM)algorithm was used to classify and recognize the five predefined arm movements.The recognition rate reached 99.33%.It verifies the feasibility of the proposed arm motion recognition method.

关 键 词：数据手套 手臂模型 动作轨迹特征 动作识别 支持向量机 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]