眼鏡式之個人化跌倒偵測機制

Self-Adaptive Fall-Detection Apparatus Embedded in Glasses

郭智榮(Chih-Jung Kuo)

指導教授 ： 陳自強

國立中正大學/工學院/電機工程學系/碩士(2014年)

摘要

跌倒造成受傷已經是老人健康照護主要問題，本研究所提的系統採用了包含三軸加速度感測器和三軸陀螺儀感測器的六軸感測模組進行偵測，主要目的是開發一個穿戴式跌倒偵測設備，將之與眼鏡相結合，以達到方便、舒適、非侵入式的便利性。
在本研究中，針對跌倒判斷提出了一個以H-SVM (Hierarchical Support Vector Machine)架構，動作姿勢通過使用六軸感測器，來進行偵測判斷，主要將動作分成兩類：動態姿勢和靜態姿勢，分別識別出四種靜態姿勢(站立、行走、上樓梯、下樓梯)與五種動態姿勢(向前傾倒、向後傾倒、右側傾倒、左側傾倒、慢跑)，並通過感測器的使用來判斷跌倒的方向性，當跌倒發生時能精準的進行警報，實驗結果顯示，本跌倒偵測系統的跌倒辨識率高達97.60%，而姿勢辨識率高達92.92%。相比於傳統的跌倒偵測系統，所提出的系統不僅顯示了較好的性能，也提供了方便，舒適，非侵入式的佩戴方式。因此，本研究所提出的系統可以應用於醫療方面居家照護，以提供使用者一個可行的穿戴式偵測設備。

關鍵字

姿勢判斷 ； H-SVM (Hierarchical Support Vector Machine) ； 六軸感測模組 ； 跌倒判斷 ； 非侵入式配戴方式

並列摘要

Fall injury is already a major problem in elderly health care. The proposed system adopts a 6-axis sensing module of a triaxial accelerometer and a gyroscope. The main purpose of this work is to develop a wearable headset fall-detection apparatus. Such an apparatus can be integrated with glasses to reach convenient, comfortable and non-intrusive utilization.
In this thesis, we present a novel fall detection system using H-SVM (Hierarchical Support Vector Machine). We partition human activities into two categories: static postures and dynamic postures. By using a 6-axis sensing module of a triaxial accelerometer and a gyroscope at the same body location, our system can recognize four kinds of static postures (standing, walking, going up stairs and going down stairs) and five kinds of dynamic postures(forward fall, backward fall, right fall, left fall and jogging). When a fall occurs, its direction is identified using a gyroscope and an accelerometer. The experimental results reveal that the proposed system achieves fall accuracy rate of 97.6%, and posture accuracy rate of 92.92%. As compared to the conventional fall-detection systems, the proposed system not only shows fairly good performance but also provides convenient, comfortable and non-intrusive wearing. Therefore, the system proposed herein can be widely spread in various head-mounted devices for health care applications.

並列關鍵字

posture-detection ； fall-detection ； H-SVM (Hierarchical Support Vector Machine) ； 6-axis sensing module ； non-intrusive wearing

參考文獻

[3] Xinguo Yu, “Approaches and Principles of Fall Detection for Elderly and Patient”, Proc. of IEEE International Conference on e-Health Networking, pp. 42-47, Jul. 2008.

[10] Jiangpeng Dai, Xiaole Bai, Zhimin Yang, Zhaohui Shen, and Dong Xuan, “Perfalld: apervasive fall detection system using mobile phones,” Proc. of IEEE International Workshops on Computing and Communications, pp. 292–297, Mar. 2010.

[12] W.-C. Cheng and D.-M. Jhan, “Triaxial Accelerometer-Based Fall Detection Method Using a Self-Constructing Cascade-AdaBoost-SVM,” Journals of IEEE Transactions on Biomedical and health informatics, vol. 17, pp. 411–419, Mar. 2013

[15] L. C. Molina, L. Belanche, and A. Nebot, “Feature selection algorithms: a survey and experimental evaluation,” Proceedings of the IEEE International Conference on Data Mining, pp. 306-313, 2002.

[16] Dean M. Karantonis, Michael R. Narayanan, Merryn Mathie, Nigel H. Lovell, and Branko G. Celler, “Implementation of a real-time human movement classifier using a triaxial accelerometer for ambulatory Monitoring,” Journals of IEEE Transactions on Information Technology in Biomedicine, vol. 10, pp. 156–167, Jan. 2006.