腕带式跌倒检测系统研究

发布时间：2018-05-14 21:41

本文选题：跌倒检测 + 能耗管理　；参考：《南京邮电大学》2016年硕士论文

【摘要】：当前,全球人口老龄化问题越来越严重。据联合国预测,2150年老龄人口将占据世界人口的30%,因此老人健康问题是当前社会面临的主要问题。随着空巢老人数目增加,老人跌倒死亡比例急剧增加,而跌倒检测可以及时发现老人跌倒事件,获得及时救助,减少死亡率。可穿戴技术出现,为跌倒监测提供了一种新的解决方法。但当前该技术在实际应用中还存在准确度低、舒适度差和设备能耗高等问题。针对上述问题,本文提出了腕戴式跌倒检测技术,利用三轴加速度传感器倾角测量机制,判断跌倒事件,提高了跌倒事件判断准确性、携带舒适性及能效性问题。并在可穿戴平台Shimmer上对算法和能效性做了验证。主要研究成果如下:(1)当前基于三轴加速度传感器的跌倒监测算法判断准确度低、携带不方便。基于存在的问题,本文提出腕带式系统,利用腕带内的三轴加速度传感器监测老人姿态,根据姿态数据梯度,选用双阈值算法,判断跌倒事件。实验中对不同分类算法进行灵敏度、特异度以及复杂度比较。结果表明,本文提出的算法在特异度相同条件下,灵敏度可以提高25%。(2)可穿戴设备中能耗问题是影响系统生命期主要因素。本文针对可穿戴平台Shimmer节点,对传感、计算以及发送模块分别进行了能耗测试,提出了事件驱动的数据传输和节点休眠机制。通过特征值在线判断,当特征值超过阈值才发送数据,否则不发送;通过倾斜与振动传感器控制节点休眠,减少系统能效性。实验结果表明,休眠机制可以延长佩戴时间从16小时到29小时,生命期延长接近80%。(3)在可穿戴节点Shimmer平台上,对提出的跌倒判断检测算法和能耗节约策略进行验证。在低功耗、非实时操作系统Tinyos环境下,发现精度缺失和数据发送丢包等问题,实际获得的灵敏度是要比理论测量值要小20%,特异度几乎没变,生命期延长30小时。本文研究工作在跌倒检测中具有一定的学术价值,在医疗、工业以及社会等领域具有较大的应用前景。
[Abstract]:At present, the global population aging problem is becoming more and more serious. According to the UN forecast, the aging population will account for 30% of the world's population in 2150, so the health of the elderly is the main problem facing the society. With the increase of the number of empty nest elderly, the death rate of the elderly tumbling increases sharply, and the fall detection can find the fall events in time, get timely rescue, and reduce the mortality rate. The appearance of wearable technology provides a new solution for fall monitoring. However, there are still some problems in practical application, such as low accuracy, poor comfort and high energy consumption. In order to solve the above problems, this paper puts forward the technique of wrist wear type fall detection. The triaxial acceleration sensor is used to judge the fall event, which can improve the accuracy, comfort and energy efficiency of the triaxial acceleration sensor. The algorithm and energy efficiency are verified on wearable platform Shimmer. The main research results are as follows: (1) the current triaxial acceleration sensor based fall detection algorithm has low accuracy and is not convenient to carry. Based on the existing problems, this paper proposes a wrist band type system, which uses the three-axis acceleration sensor in the wrist band to monitor the attitude of the elderly. According to the gradient of attitude data, the double-threshold algorithm is selected to judge the fall event. In the experiment, the sensitivity, specificity and complexity of different classification algorithms are compared. The results show that under the same specificity, the sensitivity of the proposed algorithm can be improved by 25 / 2) the energy consumption problem in wearable devices is the main factor affecting the system life cycle. In this paper, the energy consumption of sensing, computing and sending modules is tested for wearable platform Shimmer nodes, and an event-driven data transmission and node hibernation mechanism is proposed. Through the on-line judgment of the eigenvalue, the data is sent when the eigenvalue exceeds the threshold value, otherwise the data is not sent, and the energy efficiency of the system is reduced by controlling the dormancy of the node by tilting and vibration sensor. The experimental results show that the sleep mechanism can prolong the wear time from 16 hours to 29 hours and the life span by nearly 80%) on the wearable node Shimmer platform, the proposed fall judgment detection algorithm and energy saving strategy are verified. In the low power, non-real-time operating system (Tinyos) environment, the problems of missing precision and losing data packet are found. The sensitivity obtained is 20% less than the theoretical measurement value, the specificity is almost unchanged, and the life span is extended by 30 hours. The research work in this paper has certain academic value in the fall detection, and has a great application prospect in the fields of medical treatment, industry and society.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH789

【相似文献】