透過您的圖書館登入
IP:18.118.200.86
  • 學位論文

基於智慧型手機之跌倒偵測與三維計步器之設計與實現

Design and Implementation of Fall Detection and Triaxial Pedometer Based on Smartphone Platform

指導教授 : 黃連進

摘要


本文利用智慧型手機內建之三軸加速感測器和地磁感測器來計算走路步數,最終目的為老年人之跌倒偵測。高齡族群使用智慧型手機接受度提高,使用智慧型手機內建之感測器進行偵測,可降低感測器大小及配戴多種感測器之不便,以手機內建的多種感測器得知使用者肢體動態計算步數與分析跌倒狀況是否發生,防止當跌倒發生時,無人在旁協助治療之情況。 本文計步器演算法在分析三軸加速度量值於平地行走可達96%準確率,上樓梯可達90%準確率,下樓梯可達86%準確率,並可偵測使用者目前肢體動作以達更高之準確率,跌倒偵測改良Maarit Kangas等多位學者演算法,加入肢體動作分析,增加準確度避免因較激烈日常活動而產生跌倒誤判,增加使用者困擾。日常行走與起立坐下時不會有錯誤偵測跌倒警告,準確率達100%,整體跌倒偵測準確率可達88%。

並列摘要


This thesis use two sensors tri-axial acceleration sensors and magnetic sensors in smartphone to calculate walking steps. The main purpose is for detecting of fall in the elderly. Due to the older adult’s acceptance attitude of using smartphone is improving, using sensor in smartphone to detect fall will reduce incommodiousness of carry, and the volume of sensor become smaller. With various sensors to detect fall, could know that the user's calculating steps in dynamic and analyze the situation of fall will happen or not. In case the fall happened, no one help the elderly. The algorithm of pedometer analyzed the accuracy rate of three-axis acceleration, it may reach 96% in walking, 90% in going upstairs, and 86% in going downstairs, it also achieves a higher accuracy rate in the user's physical action. The fall detection improved Maarit Kangas' and other scholars' algorithm. Adding the analysis of physical action raise the accuracy to avoid misjudgment of violent daily activities and use’s inconvenience. Daily walking, Standing up and sitting down will not warn the user of fall, the accuracy rate of whole fall detection may reach 88%.

參考文獻


[20] 許輝煌、蔡康俊,“使用手機重力感應器和類神經網路於姿態辨識”,淡江大學資訊工程學系碩士班,2012。
[21] 黃有評、王珊珊,“以智慧型手機感測器設計居家照護系統”,國立台北科技大學,電機工程系所,2011。
[28] 陳瑞發、林顗德,“在智慧型手機上應用重力感測器進行人體動作分析”,淡江大學資訊工程學系碩士班,2011。
[34] 吳永基、蔡昇倫,“智慧型手機之跌倒偵測系統設計”,長榮大學,資訊管理學系碩士班,2012。
[3] 中華民國護理師護士公會全國聯合會,URL:http://www.nurse.org.tw/

被引用紀錄


Tseng, C. C. (2016). 基於影子偵測和邊線偵測之行人避障輔助系統應用於智慧型手機 [master's thesis, Chung Yuan Christian University]. Airiti Library. https://doi.org/10.6840/cycu201600028

延伸閱讀