Real time multi parameter detecting method for home health care★

Abstract

AIM： To probe into the real-time detecting method of multi-parameter Home Health Care on electrocardiogram (ECG), blood pressure and activity state.

METHODS： The signals of ECG, blood pressure and activity state collected from Daping Hospital Affiliated to Military Medical University of Chinese PLA between October 2003 and October 2004 were detected in real-time. To improve the accuracy for real-time detection of ECG, techniques of self-learning, classification detecting and adaptive parameters adjusting were adopted. Blood pressure parameters were calculated by using amplitude-coefficient method, and two ratio coefficients were obtained statistically from lots of data. MMA220 was used to catch the signal of acceleration, and the speed of which was obtained by segmental integral, so that the activity state can be deduced. All algorithms were performed by integrating C language with assembly language on MSP430 equipment.

RESULT:① The accuracy was higher than 96% in detecting ECG with the state of keeping quiet, walking and running slowly. Besides, it is not complex with small amount of calculation, which is helpful to the realization of machine. Comparison in measurement data between with mercuric sphygmomanometer and method of ratio coefficients was performed, and the average discrepancy value of systolic blood pressure (SBP) was 1.84 mm Hg and that of diastolic blood pressure (DBP) was 0.96 mm Hg. Distortion of speed caused by the error of acceleration was decreased obviously when the speed was calculated step by step. ② The algorithm consumes 3 milliseconds running in MSP430F149.

CONCLUSION：The proposed real-time detecting algorithm for ECG, blood pressure and activity state has better accuracy and lower complexity, so it is suitable for home health care.

摘要
目的：探讨多参数家庭健康监护系统终端对心电、血压及活动状态三种信号进行实时检测的方法。
方法：对解放军第三军医大学大坪医院2003－10/2004－10采集的心电、血压和活动状态三种信号并进行实时检测。对于心电信号，设计了结合自学习、分类检测、参数自适应调整的实时检测算法；血压参数采用幅度系数法检测，通过大量的实验数据，经过统计确定两个幅度系数的具体数值；采用MMA220加速度传感器获取活动时的加速度，通过分段积分的方法获得瞬时速度，再结合加速度进行活动状态判断。以上算法均采用C语言和汇编语言相结合在MSP430单片机上得以实现。
结果：①心电信号检测方法检测准确度高，正常心电在安静、慢走、慢跑等活动状态下，正确率在96%以上，并且复杂度低、运算量小，利于单片机的实现；对采用的幅度系数法进行的血压测量和汞柱血压计测量的检测结果进行统计比较，收缩压差值的平均值为1.84 mm Hg，舒张压差值的平均值为0.96 mm Hg；对于运动状态的检测，分段积分计算速度值，可以较好地减少加速度信号中存在误差造成速度曲线的失真。②检测方法在MSP430F149单片机上实现，运行时间为3 ms。
结论：多参数家庭健康监护实时检测方法准确率高，方法简单可行，能满足家庭健康监护的要求。
关键词：监护；心电描技术；血压

朱新建，吴宝明，卓豫，朱凌云.多参数家庭健康监护实时检测的方法[J].中国组织工程研究与临床康复，2007，11(1):121-123
[www.zglckf.com/zglckf/ejournal/upfiles/07-1/1k-121(ps).pdf]