本論文以智慧型手機內部加速規,接收人體行走訊號,轉換成具有碎形性質的步伐間隔時間,並且做初步驗證與分析,來判斷生理健康狀態。使用者可透過智慧型手機便可幫助健康狀態監控,而不需到醫院做昂貴的健康檢查,減少醫療資源耗費。手機優點在於隨時隨地都可量測訊號並且監控健康狀態,以手機個人化的特性,提供專屬化的管制圖監控,而達到發現不健康狀態,便可即時檢查與治療。因步伐的間隔時間訊號含有碎形的特性,本研究以去趨勢擾動分析(Detrented Fluctuation Analysis, DFA)來分析步伐的間隔時間訊號,並計算出碎形特徵值α 與建立管制圖進行監控。本研究以智慧型手機進行步伐間期分析與監控之研究成果,確立了手機加速規量測行走訊號的可靠程度,可作為未來建構步伐間期分析系統之基礎。
This thesis aims at the design and implementation of a human healthcare system using the smart phone. Smart phone with the personal and mobile property is applied to collect human physiological signals. Statistical control charts are then built on the smart phone to serve for personalized health monitoring at anytime and anywhere. Among various human physiological signals, we select the stride interval characterizing the time between successive walk steps as the problem conveyer. The data sources for stride interval calculation are the signals collected by the accelerometer embedded in smart phone. We apply the technique of Detrented Fluctuation Analysis (DFA) to analyze the signal of stride intervals and calculate the fractal property as an index to monitor the state of human health. A prototype system is implemented to demonstrate the feasibility of smart phone-based stride interval calculations for personalized health monitor.