- 學位論文
基於IoTivity框架之推論式即時人體跌倒姿勢重建
Inferential Real-time Human Fall Posture Reconstruction Based on IoTivity Framework
摘要
本論文實做一個即時人體跌倒姿勢重建的系統，透過三軸加速度計與三軸陀螺儀來偵測跌倒的發生，並利用姿態演算法來估算人體各部位角度，利用無線人體感測網路(Wireless Body Sensor Network)來收集各部位感測器資訊，結合IoTivity的開源物聯網協定，來架構使用者與設備之間的溝通，所有資訊將會透過IoTivity轉為資源供使用者查閱，並在客戶端使用3D Model的方式來重現人體跌倒姿勢，使跌倒流程更為直觀地顯示。本研究透過貝式網路來推論下一個動作的發生，推論式的貝氏機率表能夠提醒醫護人員更為完整的跌倒過程資料，即使數據資料因為傳輸網路或設備受損也可以利用貝氏機率表來推論下一個姿態，作為急救的參考。
並列摘要
This paper implemented the real-time reconstruction of human fall-down posture. Through three-axis accelerometer and three-axis gyroscope to detect the fall-down event, and use the attitude algorithm to estimate the angle of the human body. Using wireless body sensor network to collect the various parts of the sensor information, combined with the open source agreement of Internet of Things about IoTivity, to structure communication between user and device. All of the information will be converted to resource for users through IoTivity, and reconstructed human fall-down posture by 3D model at client. Make the fall-down process more intuitive. In this study, through Bayesian networks to infer the next action. Inferential Bayesian probability table can remind health care worker more complete information on the process of fall-down. Even if the data transmission network problem or device damage can also use inferential Bayesian probability table to the next posture as a reference aid.