基于惯性传感器的可穿戴式帕金森震颤检测系统设计  

Design of Wearable Parkinson’s Tremor Detection System Based on Inertial Sensor

作　　者：佟丽娜[1] 刘岱松 何佳纪 张明佳 彭亮[2] TONG Lina;LIU Daisong;HE Jiaji;ZHANG Mingjia;PENG Liang(School of Mechanical Electronic&Information Engineering,China University of Mining and Technology-Beijing,Beijing 100083,China;State Key Laboratory of Management and Control for Complex Systems,Institute of Automation,Chinese Academy of Sciences,Beijing 100190,China)

机构地区：[1]中国矿业大学(北京)机电与信息工程学院,北京100083 [2]中国科学院自动化研究所复杂系统管理与控制国家重点实验室,北京100190

出　　处：《传感技术学报》2023年第2期314-322,共9页Chinese Journal of Sensors and Actuators

基　　金：北京市自然科学基金项目(Z170003,4232053)。

摘　　要：为了辅助帕金森震颤诊疗,设计了基于惯性传感器的可穿戴式帕金森震颤检测系统。在数据采集、Bluetooth通讯、电源管理和角度解算等方面做了详细设计,并在手机端设计实现了数据接收及震颤识别APP,利用UCOS实现了对硬件断线重连和数据同步接收,对装置各种性能进行测试,设备续航时间可达17 h,运动数据的丢包率低于0.2%,角度误差在±1°范围内,同时提出了基于CNN的帕金森震颤识别方法,准确率和特异性分别为97.3%和97.2%。系统的各项测试结果均满足实际医疗使用需求,为帕金森肢体震颤辅助诊疗提出了一个可行的方法。In order to assist the diagnosis and treatment of Parkinson tremor,an inertial sensor based wearable detection system for Parkinson’s limb tremor is designed.The components of data acquisition,bluetooth networking,power management and angle calculation are designed in detail.At the same time,an APP for data reception and tremor recording on the mobile phone is designed.UCOS is used to realize hardware disconnection reconnecting and data synchronization reception.The performance test shows that it has a battery life of 17 hours,the packet loss rate of sports data is lower than 0.2%,and the angle error is within±1°.At last,a CNN-based recognition algorithm for Parkinson’s tremor is proposed,the accuracy and specificity are 97.3%and 97.2%.The test results of the device meet the actual medical needs.It provides a feasible method for the assist of diagnosis and treatment for Parkinson’s limb tremor.

关 键 词：帕金森病 震颤检测 惯性传感器 蓝牙组网 数据同步接收机制 卷积神经网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]