運用建築資訊模型為基礎之室內定位與維修系統-以室內停車場為例

Abstract

近年來，由於建築資訊模型(Building Information Modeling, BIM)的崛起以及行動通訊的盛行與普及，室內定位系統的服務與維修特別受到重視。然而室內定位因為建築與結構的環境限制，面臨重大挑戰。為了因應此問題，本研究旨在建構一個創新且完整的室內定位與維修系統，所提出的系統運用是以建築資訊模型為基礎的無線網絡佈置(BIM based Access Point Deployment, BAPD)，來進行室內定位與坑洞偵測。以工業基礎分類(Industry Foundation Classes, IFC)的文件作為輸入及輸出來建構一新標準資料交換格式演算法IFA，在室內定位，BAPD可以利用最少的無線網路佈點(IEEE 802.11)，提供最完整的信號覆蓋，並將蒐集到的信號強度、傳送時間和傳輸距離均列入定位考量依據。 此外，在建築樓面維護的部分，針對坑洞偵測提出建議方案，該方法首先透過行動通訊數據分析，找到潛在的坑洞，再利用空間內插法來識別坑洞的確切位置。在實作的過程中，並使用擴增實境(Augmented-Reality)技術開發系統，提供有效並且有用的系統運用。研究結果顯示IFA 整體正確率達94.96%，BAPD有效降低佈點點數達29.27%，室內定位整合異質網路可將誤差從15.89公尺降低至79公分，且即時坑洞偵測之誤判率為零，最後，運用擴增實境能夠有效指引使用者最近的停車位並標示方向與距離。

For the past few years, indoor positioning and maintaining has been drawn much attention. However, because of environmental limitation and building structure, it is still a challenging problem. To address this problem, in this thesis, we propose a state-of-the-art indoor positioning and maintaining system. The proposed system supports a Building Information Modeling (BIM) based Access Point Deployment (BAPD), indoor positioning and pothole detection. By using the BIM/Industry Foundation Classes (IFC) file as input/output to form a new algorithm IFA, the BAPD can identify the minimal set of access points (IEEE 802.11) while still providing full signal coverage. As to indoor positioning, signals of both cellular networks (4G/3G) and local area networks (IEEE 802.11) are examined. Signal strength, time serial, and transmission characteristic are taken into consideration for positioning. Also, for maintaining building floors, a pothole detection method is suggested. This method first finds potential potholes by analyzing mobile sensor data. It then tries to identify the exact location of the potholes by using the spatial interpolation algorithm. In practice, the augmented-reality technology is used to develop the system. Simulation results indicate the proposed system is both effective and efficient. The results show that IFA overall accuracy rate is 94.96%, BAPD effectively reduces the distribution access points down to 29.27%, the integration of indoor positioning with heterogeneous network may be reduced the tolerance of positioning from 15.89 meters to 79 centimeters, and the error rate of the real-time pothole detection is zero. Finally, the use of augmented reality could effectively direct the user go to the nearest parking space and mark the direction and distance.