基于状态识别的人体运动中穿戴设备动态监控系统  

Dynamic monitoring system of wearable devices in human motion based on state recognition

作　　者：马庆[1] 李杰[1] 王远超[1] 许崇高[1] MA Qing;LI Jie;WANG Yuanchao;XU Chonggao(Xi’An Fanyi University,Xi’an 710105,China)

机构地区：[1]西安翻译学院,西安710105

出　　处：《自动化与仪器仪表》2023年第1期166-171,共6页Automation & Instrumentation

基　　金：陕西省科技厅面上项目《基于虚拟条件下高尔夫全挥杆动作学习的反馈控制以及技能迁移研究》(2022JM-138);西安翻译学院《大学生体育课程综合改革理论与实践创新研究》(XFU20KYTDC02)。

摘　　要：针对运动动作采集与识别的需求,提出一种基于IMU数据采集与SVM的动作识别的可穿戴动态监控系统。在该系统中,首先利用IMU采集运动数据,然后对数据进行平滑滤波处理,以消除噪声对数据的影响,动作特征的提取将利用动作分割算法,动作的识别采用支持向量机。结果证明:SVM算法在经过特征选择后,在动态动作、静态动作以及全部动作的识别率上达到了98.01%、99.23%、98.52%,整体识别效果好,证明了基于状态识别的可穿戴设备监控系统具有可行性。Aiming at the problem that vision based motion recognition system is greatly affected by the environment, and the rapid development of wireless transmission technology has laid the foundation for wearable device based human motion recognition system, a wearable device dynamic monitoring system based on state recognition is proposed. By placing IMU in the appropriate part of the body, and then collecting the data output by the sensor during movement;Then the data is smoothed and filtered to eliminate the influence of external noise on the data;An action segmentation algorithm based on threshold is designed to extract useful action segments;Finally, support vector machine is used to recognize actions. The experimental results show that the SVM algorithm has a good overall recognition effect after feature selection, which proves that the wearable device monitoring system based on state recognition is feasible.

关 键 词：动作识别 特征提取 决策树 支持向量机 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]