摘要
目的探索一种基于惯性传感器信号进行上肢康复运动基本动作单元的识别方法,以实现脑卒中患者的无监督锻炼。方法对上肢康复锻炼中的6种基本动作单元进行数据采集和识别:利用惯性传感器采集上肢运动数据,进行特征提取使用机器学习模型进行分类。结果采用极限学习机、支持向量机、随机森林和k近邻四种分类器对实验采集的6种基本动作单元的平均识别率可达到99%左右,误差不超过1%。结论本研究提出的基于惯性传感器进行脑卒中上肢基本康复动作识别方法具有一定的有效性和鲁棒性。
Objective To explore a recognition method of basic action units of upper limb rehabilitation based on inertial sensor signals,which can realize unsupervised exercise for stroke patients.Methods Data collection and recognition were carried out for six basic action units in upper limb rehabilitation exercise.Firstly,inertial sensors were used to collect upper limb motion data.Secondly,feature extraction was carried out to extract the statistical features in the data.Finally,machine learning model was used for classification.Results Using extreme learning machine,support vector machine,random forest and k-nearest neighbor,the average recognition rate of six basic action units collected in the experiment can reach about 99%and the error is no more than 1%.Conclusion In this study,the proposed method based on inertial sensor for basic rehabilitation of upper limb after stroke has certain effectiveness and robustness.
作者
辛在海
邢蒙蒙
曹慧
杨锋
魏稣濛
XIN Zaihai;XING Mengmeng;CAO Hui;YANG Feng;WEI Sumeng(College of Intelligence and Information Engineering,Shandong University of Traditional Chinese Medicine,Jinan Shangdong 250355,China;China Rehabilitation Research Center,Beijing 100071,China;Department of Equipment,Affiliated Hospital of Shandong University of Traditional Chinese Medicine,Jinan Shandong 250000,China)
出处
《中国医疗设备》
2022年第5期33-36,共4页
China Medical Devices
基金
国家自然科学基金项目(81973981)。
关键词
惯性传感器
康复动作识别
基本动作单元
动作采集
脑卒中
inertial sensor
rehabilitation action recognition
basic action unit
motion collection
cerebral stroke
