應用支援向量機於日常活動之判斷

Identifying Activities of Daily Living Based on Support Vector Machine

梁議中(Yi-Chung Liang)

指導教授 ： 方士豪

元智大學/電機通訊學院/電機工程學系/碩士(2012年)

摘要

智慧型行動裝置發展至今，由於運算能力的提升以及內建各式各樣的感測器，使得智慧型手機更貼近我們的生活。藉由智慧型手機是有可能改變現在的醫療看護方式，其中姿態偵測就是一個可能發展的項目。在日常生活中，因為跌倒而受傷的人相當多，尤其是老人，除了會下降生活品質外，嚴重甚至會導致死亡。本研究提出一個在Android平台上的跌倒偵測系統，主要分成三個部分，加速度感測、學習行為間的關係以及緊急通報。使用內建的感測器截取加速度資料，再透過單一門檻值進行二元分類(Binary-class)判斷是否跌倒。接著，我們利用相同的理論變成多元分類(Multi-class)，使用機器學習理論中的支援向量機(SupportVector Machine)來判斷日常生活的行為(Activities of Daily Living)。不同於以往使用門檻值的方法，將智慧型手機配戴在四個人的胸部、腰部和大腿上收集加速度資料，並標記各個日常生活行為(例如: 走路、跑步、坐下和躺下)的類別，再透過核心函數(Kernel Function)將訓練資料映射到更高維的特徵空間(Feature Space)分類。實驗結果顯示此方法能有效提升系統偵測跌倒的能力，並且判斷日常生活行為。除此之外，本論文也討論了不同因素的影響，例如配戴的位置和量測的數量，包括加速度、角度或是時間長度(Temporal)。我們也透過ROC(ReceiverOperating Characteristic)曲線觀察不同核心函數對系統的影響。

關鍵字

跌倒偵測 ； 行動裝置 ； 加速度感測計 ； 支援向量機

並列摘要

Since todays smartphones are programmable and embed various sensors, these phones have the potential to change the way how healthcare is delivered. Fall detection is definitely one of the possibilities. Injuries due to falls are dangerous, especially for elderly people, diminishing the quality of life or even resulting in death. This study presents the implementation of a fall detection prototype for the Android-based platform. The proposed system has three components: sensing the accelerometer data from the mobile embedded sensors, learning the relationship between the fall behavior and the collected data, and alerting preconfigured contacts through message while detecting fall. Unlike traditional methods which identify falls based on thresholds, the proposed algorithm utilizes support vector machine to learn different activities of daily living (ADL). We collect real accelerometer data from mobile embedded sensors and label the corresponding user behavior, including sitting, walking, lying, and running. Experimental results show that the proposed system outperforms traditional threshold-based scheme, increasing the fall detection accuracy by 10%. The experiment also investigates the impact of different factors, such as the locations where the device attached and the numbers of measurements, including the acceleration, angle variation, and temporal length. We also investigate the effects of different kernel functions to the proposed mechanism through receiver operating characteristic.

並列關鍵字

fall detection ； mobile phones ； accelerometer ； support vector machine

參考文獻

[5] W.-Y. Shieh and J.-C. Huang, "Speedup the multi-camera video-surveillance system for elder falling detection," in International Conference on Embedded Software and Systems, pp. 350-355, 2009.

[6] C.-W. Lin and Z.-H. Ling, "Automatic fall incident detection in compressed video for intelligent homecare," in International Conference on Computer Communications and Networks, pp. 1172-1177, 2007.

[7] F. Sposaro and G. Tyson, "ifall: An android application for fall monitoring and response," in International Conference of the Engineering in Medicine and Biology Society, pp. 6119-6122, 2009.

oor-vibration based fall detector for elderly," in Information and Communication Technologies, vol. 1, pp. 1003-1007, 2006.

[10] D. Karantonis, M. Narayanan, M. Mathie, N. Lovell, and B. Celler, "Implementation of a real-time human movement classi er using a triaxial accelerometer for ambulatory monitoring," IEEE Transactions on Information Technology in Biomedicine, vol. 10, no. 1, pp. 156-167, 2006.