實境導航系統之研究

Research of Live-view Navigation System

李鎰帆(Yi-Fan Li)

指導教授 ： 黃俊堯

國立臺北大學/電機資訊學院/通訊工程研究所/碩士(2010年)

摘要

本研究針對傳統導航系統問題，運用周遭感知(context-aware)技術，提出一套智慧型導航技術。傳統導航系統主要是運用GPS定位系統結合地理資訊系統執行路徑規劃，並提供使用者即時導航指引服務，然而傳統導航系統卻常面臨GPS硬體誤差與人為判斷誤差等兩大類問題，造成導航效果大打折扣，甚至發生意外的情形。傳統在解決GPS硬體誤差方面的研究，過去所仰賴的是絕對定位與相對定位技術，絕對定位技術是指如獨立站台等高成本技術，相對技術則是採用輔助慣性設備或是Map Matching技術；而在解決人為判斷誤差方面，由於無法掌握使用者導航行進意圖，現今的導航系統都僅是確認使用者是否在規劃路徑上，如果沒有則提供迴轉指示，要不然就是重新規劃路徑，常讓使用者陷入混淆的狀況。本研究以普及式運算為出發點，提出一套以人為本的智慧型導航服務，所謂的智慧型就是藉由探討使用者於導航過程針對規劃路徑所反應的行為習性來提供適性的導航服務。於導航過程，當使用者未按照規劃路徑行進時，本研究以周遭感知技術推論出使用者是否正處於迷路狀態，如果是，則進一步提供迴轉訊息，如果不是，系統會以使用者的行為習性分別就GPS硬體誤差與人為誤差問題進行自我調整。所提出的自我調整解決方案，係以周遭感知技術並擷取位置與方位感測器資訊，在訓練階段中，運用隱藏式馬可夫模式與牛頓運動定律將使用者穩定行為習性紀錄在轉換機率矩陣中，使得在導航階段，當導航系統偵測出使用者行進路徑與系統規劃路徑不一致時，可透過轉換機率矩陣及系統規劃路徑來推論使用者的行為狀態，據以判斷出此不一致情形是屬於導航系統的硬體誤差問題抑或是使用者的行為問題。針對GPS硬體誤差，本研究能透過周遭感知技術予以修正；而針對使用者行為問題，本研究透過使用者的移動擺盪狀態來預測使用者所面臨的下一個中途點(waypoint)，以提早重新規劃路徑的時機，並提供無縫隙的智慧型導航服務。最後，本研究提出的方法除了可以解決2009年AR street view所提出的靜態實景系統所面臨的Position Inaccuracy與Orientation Inaccuracy問題之外，並透過行動擴增實境(Augmented Reality，AR)技術提供使用者即時的動態實景導航圖示，模擬驗證所提出的智慧型導航技術之可行性，並以一連串的實作實際探討在不同環境下，所提出之實景導航系統與傳統導航系統的導航習性差異，所提出之實景導航感知技術，將可大幅縮短提供使用者路徑規劃的時間並較傳統衛星導航系統提供更為適時、直覺與便利的導航服務。

關鍵字

周遭感知 ； 擴增實境 ； 智慧型手機 ； GPS導航 ； 行動導覽

並列摘要

With the rapid development of embedded technology, GPS navigation devices become cheaper yet more powerful. Most users are often confused by the difference between environmental recognitions and geographical navigation data. Since the position and orientation information are received from GPS sensor which is easily influenced by environmental interference, it will further reduce the effect of GPS navigation. This paper presents the smart perceptive waypoint aware service of legacy GPS navigation system to provide seamless Location Based Service to solve the interruptive and interference problem of navigation service. The proposed real-time navigation technique is based on context-aware techniques to solve the re-route planning problem, which can lead the user to have disorientation situation. For example, since the legacy GPS navigation system cannot acquire the navigation status of user and then provide the wrong path to the user. This paper proposes a novel approach, called Waypoint Service, to improve the timing of re-route planning accuracy directly from the contextual information of received user behavior data. To utilize the characteristics of navigation sequential and temporal data, Waypoint aware starts by sample received navigation data into motion states according to Newton’s Law of Motion and derive Newton Markov Model (NMM) to describe the relationship among these states. Based on NMM, Waypoint aware captures the carrier’s behavior into a stable Transition Probability Matrix (TPM) during the training phase. Using the stable TPM, Waypoint aware infers the abnormal behavior of the GPS carrier from online received navigation data and offer the next reasonable waypoint. Based on PGPS technique, Waypoint aware is designed to perceive the relationship between user’s movement and his surrounding geographical environment from the integration of GPS digital compass on the Android power platform. The Augmented Navigation service is further integrated to provide intuitive navigation interface to the user. The study conducts the serious of experiments to verify the feasibility of proposed waypoint aware service on the common carrier. The experiment results reveal the waypoint aware can offer the earlier timing of the re-route planning than traditional GPS navigation system. The proposed waypoint aware service can provide more timely, intuitive and convenient navigation service.

並列關鍵字

Context-aware ； Augmented Reality ； Smart Phone ； GPS navigation ； Mobile Guide

參考文獻

[4] Jiung-yao Huang, Chung-Hsien Tsai, Shing-Tsaan Huang, “PGPS: A Positioning Technique by Perceiving GPS Data”, submitted to Communication of ACM, Nov. 14, 2009.

[5] Michael D. McDaniel (2010). Agent-Based Modeling of Lost Person Wayfinding. Master' s Thesis. University of Santa Barbara.

[8] Weiser, M., The computer for the twenty-first century, Scientific American, September, 1991, pp. 94–10.

[9] M. Satyanarayanan, Pervasive computing: Vision and challenges, IEEE Pers. Comm. 8, 4 (Aug. 2001), 10--17.

[11] Borriello, G., The challenges to invisible computing, IEEE Computer, 33(11), 23–125, 2000.