老人跌倒姿态实时识别系统设计  

Design of real-time recognition system for elderly fall posture

作　　者：徐涛 孙威 色海锋 卢少微[2] 王晓强[2] 马克明[2] XU Tao;SUN Wei;SE Haifeng;LU Shaowei;WANG Xiaoqiang;MA Keming(School of Automation,Shenyang University of Aeronautics and Astronautics,Shenyang 110136,China;School of Aeronautics,Shenyang University of Aeronautics and Astronautics,Shenyang 110136,China)

机构地区：[1]沈阳航空航天大学自动化学院,辽宁沈阳110136 [2]沈阳航空航天大学航空航天学部,辽宁沈阳110136

出　　处：《传感器与微系统》2020年第3期94-97,共4页Transducer and Microsystem Technologies

摘　　要：为了实时检测老人跌倒姿态,设计了一种基于石墨烯/橡胶传感器与MPU6050的老人跌倒姿态实时识别系统,系统以MSP430F149单片机为核心,分别采集石墨烯/橡胶传感器与MPU6050的输出信号,并且通过跌倒监测算法判断是否出现跌倒。当监测到跌倒信息或者报警信息时,系统报警,同时手机端将老人的位置信息连同跌倒姿态以短信的形式发送至预设手机上,实现与急救中心以及用户家人之间的信息交互,组成一个监测网络。实验结果表明:系统监测跌倒的准确率达95.84%,跌倒姿态的准确率达90.84%,能够实现对老人跌倒姿态的实时识别,区分了6种日常状态和6种跌倒姿态。In order to real-time monitor on elderly fall posture,a real-time elderly fall posture recognition system based on graphene/rubber sensor and MPU6050 is designed,the MSP430F149 microcontroller unit(MCU)is used as the core of the system,which collects the output signals of the graphene/rubber sensor and the MPU6050 respectively,and judges a fall through the fall monitoring algorithm.The system reminds the danger when the fall information or the alarm information is detected.And the falling information and location is transmitted to the specified mobile phone by short message,to establish a monitoring network for the emergency center and the user’s family member.The results show that accuracy of the monitoring falls is 95.84%,and the identifying accuracy of the falling posture is 90.84%,which can realize the real-time fall posture recognition of the elderly,and distinguish 6 kinds of daily state and 6 kinds of fall postures.

关 键 词：跌倒姿态识别 石墨烯/橡胶传感器 MPU6050 MSP430F149 监测网络 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]