摘要
根据老年人跌倒发生率高、损伤大、抢救率低等特点,结合可计算射频识别标签(CRFID)与模式识别技术,设计一个改进的跌倒检测系统。通过CRFID采集状态信号,由经验模态分解获取信号的时频特征,对该特征进行主成分分析降维,并使用随机森林算法检测跌倒状况。实验结果表明,该系统的查全率、精确度、转移性和准确率分别为97.75%、97.9%、98%、97.8%,能实时、准确地检测老年人的跌倒行为。
According to the characteristics of the fall among elder,such as high incidence,great injury,and low rescue rate,combining Computer Radio Frequency Identification(CRFID) and pattern recognition technology,this paper designs an improved fall detection system.Firstly,the system collects the state signals by CRFID.Secondly,the state signals are decomposed by Empirical Mode Decomposition(EMD) to obtain the time-frequency features.Thirdly,Principal Component Analysis(PCA) is used to reduce the dimension of the features.Finally,the Random Forest(RF) is used to detect the falling condition.Experimental results show that the recall,precision,transferability and accuracy of the system are 97.75%,97.9%,98%,and 97.8%,respectively.The system can detect the fall behavior of the elderly accurately in real time.
引文
[1] 王田田,郭爱敏.老年人跌倒恐惧的研究进展[J].中国护理管理,2017,17(9):1217-1221.
[2] 张敏,石倩倩,张珊珊,等.基于语音识别和STM32的老年人健康状况监护系统的设计[J].电子质量,2017(5):57-59.
[3] MAGILL S S,EDWARDS J R,BAMBERG W,et al.Multistate point-prevalence survey of healthcare-associated infections[J].New England Journal of Medicine,2014,370(13):1198-1208.
[4] MUBASHIR M,SEED L.A survey on fall detection:principles and approaches[J].Neurocomputing,2013,100:144-152.
[5] BUI N N,KIM J Y,TRINHT D.A non-linear GMM KL and GUMI kernel for SVM using GMM-UBM supervector in home acoustic event classification[J].IEEE Transactions on Fundamentals of Electronics,Communications and Computer Sciences,2014,97(8):1791-1794.
[6] BELSHAW M.Towards a single sensor passive solution for automated fall detection[C]//Proceedings of IEEE Annual International Conference of Engineering in Medicine and Biology Society.Washington D.C.,USA:IEEE Press,2011:1773-1776.
[7] ELHAMOD M,LEVINE M D.Automated real-time detection of potentially suspicious behavior in public transport areas[J].IEEE Transactions on Intelligent Transportation Systems,2013,14(2):688-699.
[8] DROGHINI D,FERRETTI D,PRINCIPI E,et al.A combined one-class SVM and template-matching approach for user-aided human fall detection by means of floor acoustic features[J/OL].[2018-02-27].https://doi.org/10.1155/2017/1512670.
[9] ARIANI A,REDMOND S J.Software simulation of unobtrusive wireless sensor network for night-time falls detection[C]//Proceedings of IEEE Annual International Conference of Engineering in Medicine and Biology Society.Washington D.C.,USA:IEEE Press,2010:2115-2118.
[10] KAPOOR U,NARAYANAN M.Design of an unobtrusive wireless sensor network for nighttime falls detection[C]//Proceedings of IEEE Annual International Conference of Engineering in Medicine and Biology Society.Washington D.C.,USA:IEEE Press,2011:5275-5278.
[11] 马琼华,王忠,郑晓彬.基于Android的老年人跌倒检测系统研究[J].数字技术与应用,2017(2):105-106.
[12] 孙荣霞,马少卿,王硕南,等.一种适用于单片机的老人跌倒检测算法研究[J].现代电子技术,2017,40(22):14-18.
[13] OLIVIERI D N,CONDE I G,SOBRINOX A V.Eigenspace-based fall detection and activity recognition from motion templates and machine learning[J].Expert Systems with Applications,2012,39(5):5935-5945.
[14] 周文谊,王吉源.一种模糊森林学习方法及其行人检测应用[J].计算机工程,2017,43(3):304-308.
[15] GEORGOULAS C,BOCK T.Design and implementation of a novel cost-effective fall detection and intervention system for independent living based on wireless sensor network technologies[C]//Proceedings of International Symposium on Automation and Robotics in Construction.Pittsburgh,USA:IAARC Publications,2015:1-9.
[16] WANG Changhong,LU We.A low-power fall detector using triaxial accelerometry and barometric pressure sensing[J].IEEE Transactions on Industrial Informatics,2016,12(6):2302-2311.