可穿戴式身體運動與姿勢監控系統

Abstract

在個人健康照護的研究中，行動照護(mHealth)是電子健康照護領域中成長最快的領域。配帶這些行動看護設備(mHealth devices)使得醫護人員不只可以隨時監控患者病情，還能即時提供建議或提醒。而微型生醫感測器技術的發展與進步、以及智慧型手機、行動應用程式的出現，使得這些都變得可能。在本篇論文中提出了一個可穿戴的身體動作和姿勢監控系統並應用在肥胖患者的久坐行為識別上並幫助殘疾患者恢復健康。

我們目標是設計一個微型高效率、功能強大且可程式化的三維位置及方向感測器，它可以透過行動應用程式與智慧型手機等行動裝置溝通。此整合感測器(WIS) 可測量高達10個自由度，包含三維線性加速度、角加速度、磁場的強度及氣壓高度，取樣頻率可達100Hz且可直接處理原始資料及偵測特定的運動模式。經過處理後的訊息可經由無線整合感測閘道器(WIS&G)將訊息傳送至行動裝置。這個設備使用Squawk JVM做為底層系統環境，因此整個系統使用Java語言開發。此系統的硬體及軟體(包含無線整合感測閘道器, WIS&G)已完成開發並進行測試。

本系統的首要應用是用來識別久坐行為模式。實驗中使用一個WIS&G並配戴於上衣胸口的位置，以及一個WIS配戴於大腿位置，透過藍芽無線電連結且傳送訊號至使用者的手機，並從這些實驗中獲得初步的結果。

As an approach to ubiquitous personal healthcare, mobile healthcare (or mHealth) is the fastest growing field in electronic healthcare. With the use of mHealth devices, health professionals can not only monitor a patient’s condition as she leads her daily life, they can also offer advice or issue alerts in real time. All these have been made possible by the advent of miniature biomedical sensors, smart phones and mobile applications. In this thesis, we present the design and implementation of a wearable body motion and posture monitoring system that can be used among assorted applications to identify the sedentary behaviors of obese patients or aid the rehabilitation of handicapped patients.

Our aim was to design a miniature energy-efficient yet powerful and programmable three-dimensional position/orientation sensor that can work with a smart phone through a mobile application. This integrated sensor (WIS) is capable of gathering motion information in ten degrees of freedom (10 DOFs) including linear accelerations, angular accelerations, magnetic flux and barometric pressure up to a hundred samples per second. It can process the raw data on board to deduce real-time trajectories or detect specific motion patterns. The processed information can then be sent to a mobile device directly through its ZigBee radio or indirectly through a wireless integrated sensor & gateway (WIS&G). A Java virtual machine known as Sqwauk was installed on this device; hence, the entire system is pro¬grammable in Java. Both hardware and software of this entire system (including the WIS&G) has been developed and tested in this project.

The first application of this system was to identify the sedentary behavioral patterns. A WIS was mounted on thigh and a WIS&G clipped to the subject’s upper garment was used with a cell phone attached to her. Experiments have been carried out and preliminary results were obtained from these experiments.