請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/44548
標題: | 以人體運動特徵為基礎之精確人員定位追蹤技術研究 Accurate Human Positioning Based on Human Movement Characteristics |
作者: | Yu-Chi Lin 林郁埼 |
指導教授: | 呂東武(Tung-Wu Lu) |
關鍵字: | 慣性定位系統,人體運動數學模型,動作分類,圖資化,定位, IMUs,human model,movement classification,map,calibration,positioning, |
出版年 : | 2009 |
學位: | 碩士 |
摘要: | 行動裝置位置導向之真實情境展現界面,為一人性化的人機界面,但卻十分仰賴足夠精確之位置及姿態資訊。利用微型慣性定位技術可提供行動裝置所需的狀態資訊,並可在外部定位(如GPS、WiFi定位等)無法持續取得訊號的情況下,提供連續位置及姿態資訊,但卻有隨時間增加之累積誤差問題。
本研究旨在開發新型室內定位技術,透過人體運動特徵(human movement characteristics)建立人體於室內常見動作之人體運動數學模型。當長時間無法接收外部定位訊號,如處於騎樓或室內時,能藉由開發所得之人體運動數學模型,配合慣性定位系統所得之資料進行人體定位。本研究針對室內常見動作,包括步行、上下樓梯、坐到站以及站到坐,建立不同動作之人體運動數學模型,並透過小波轉換技術將以上動作進行判別與分類。於步行及上下樓梯之運動數學模型中,亦將轉彎列入考慮,讓此定位技術能夠更完善的運用於人體定位。 於基礎定位技術開發完成後,為求完整的定位效果與將來實用性考量,本研究將初步定位結果加入圖資,且於圖資化過程中加入座標系統,標示出定位位置,並於地圖下方顯示定位座標值,完成精確人體定位技術開發。 為測試此技術的發展性,本研究將感測器放置於人體不同位置,討論其定位效果,以及定位設備最小需求。除此之外,為驗證此技術之重複性和可靠度,透過六位不同受試者進行多項定位結果分析並進行討論。研究結果顯示,於六位不同受試者的實驗驗證下,平地走路的誤差率在3%以下,各種室內常見動作之判別成功率都在75%以上,而此技術感測器放置位置為胸骨劍突處及小腿處精準度最佳。 Location-Based Life-Reality (LBLR) realized on mobile device is proved to be a user-friendly man-machine interface (MMI). However, it heavily relies on the fact that the system must continuously get accurate position and attitude information. Inertia positioning of micro sensors can continuously provide LBLR mobile device with needed state information. However, it has the problem that its integrator accumulates the sensor errors and its accuracy dramatically decays with time. The purpose of the current study was developing the new indoor positioning technology which was founded on human mathematical model constructed with human movement characteristic during common in indoor movement. When the signal of external positioning system (EPS) is unavailable, such as people under overhang of a building or indoors, our new positioning technology which was utilized the combination of the developed human movement mathematical models and the inertial data can effectively position the human location. This study aimed to develop different movement of human movement mathematical model during common indoor movements, including level walking, stair ascent and descent, sit-to-stand and stand-to-sit, and to apply the wavelet transformation technology in reorganization and classification of all the movements. For the human mathematical model during level walking, stair ascent and descent movements, movement of turning would be also considered in the establishment of this positioning technology for more accurate human positioning. On the strength of the consideration of the positioning effects and practicability of our newly developed positioning technology, we added the map and the coordinate system in the mapping process for displaying the positioning result and the coordinates below the map shown in the monitor of the mobile positioning device. In order to test the practicability of this positioning technology, we put the inertial sensors on different body positions of human for confirming the positioning effects and the minimum requirements of this technique. Otherwise, six subjects were recruited to participate in different motion experiments for confirming the reproduction and validity of the new positioning technology. The results of different motion experiments in the current study showed that the positioning error during level walking was below 3%, the rate of success for classifying different indoor movements was greater than 75%, and the best sensor placements were Xphoid process and shank. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/44548 |
全文授權: | 有償授權 |
顯示於系所單位: | 醫學工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-98-1.pdf 目前未授權公開取用 | 3.73 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。