本研究針對傳統導航系統問題，運用普及式計算的周遭感知技術，提出新一代GPS導航技術。因為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，來提供使用者即時導航指引服務。然而傳統導航系統常面臨GPS硬體誤差、人為判斷誤差與迷航等三大類問題，造成導航效果大打折扣，甚至發生意外的情形。而傳統研究在解決這些問題方面，過去所仰賴的是絕對定位的高成本建置技術與可能產生累積誤差的相對定位技術來解決GPS硬體誤差，及藉由增加3D虛擬畫面的多畫面模式來補強道路資訊以降低人為誤差，而迷航問題則是透過提供一段預設時間的迴轉訊息之後重新路徑規畫服務來解決。而這些方式皆忽略GPS字串中所隱含的使用者行為資訊，因此本論文透過探索GPS字串內的內容資訊，應用感知用路人導航行為的技術，提出下一代導航系統，也就是實景導航系統。本論文整合感知使用者行為動向的周遭感知技術與擴增實境技術，提供以人為本的導航經驗；在所提出的感知技術中，包含用來解決導航衛星定位誤差、電子羅盤累積誤差及其所處環境干擾的位置感知與方位感知服務；還有解決迷航問題的導航點服務，最後整合擴增實境介面，消弭人為判斷誤差等問題。 本論文所提出的位置感知服務，取樣所接收的GPS字串，並依據牛頓運動定律區分成七種運動狀態，並進一步推導出牛頓馬可夫模式，配合訓練階段用路人的樣本序列所建構出穩定機率轉換矩陣，即可針對所接收的GPS字串進行用路運動狀態的合理推論，即時修正用路人的位置資訊；以及在方位感知服務上，藉由電子羅盤的輔助，交互校正移動時用路人的方位資訊；接著，配合合理的位置與方位資訊，導航點服務得以針對未按照規畫路徑的用路人行為進行推論用路人是否迷路，以縮短用路人迷航時間，最後整合擴增實境技術，提供直覺的導航介面解決人為誤差判斷。本論文藉由上述所提服務的整合，實景導航系統得以確保用路人的心智地圖與導航地圖一致。 本論文研究除了實作所提之實景導航系統，並以一連串的實驗驗證所提出的位置感知、方位感知與導航點感知技術的感知能力與可用性，並探討相關系統效能，研究中也以問卷調查來探討在不同環境下，所提出之實景導航系統與傳統導航系統的導航習性差異，結果顯示所提出之實景導航感知技術能提供使用者更為即時、直覺與便利的服務，因此較傳統衛星導航系統更易受到用路人的青睞。 Targeting on the problems of the legacy GPS navigating system, the dissertation employs the context-aware service to design the next generation GPS navigating system. The legacy GPS navigation system is frequently persecuted with GPS ranging errors, human cognitive errors, and disorientation problems. Such types of errors always lead to significant inaccurate navigation results; even more induce accidents to the user. Most of the previous efforts on solving those issues are focused on absolute and relative approaches for GPS ranging errors, use multi-window design to display 2D geographical data and simulated 3D landscape image to reduce human cognitive error, and through a preset time interval of U-turn warning before provide re-planning services. But these services ignore the contextual information of GPS data completely. To explore the contextual information of GPS data for the user’s navigational behavior, the dissertation presents the next generation GPS navigation system, i.e. a Live-view GPS navigation system, which integrates context-aware techniques to perceive user’s intention from continuous sensors data with Augmented Reality technologies to provide a human-centric navigating experience. The perceptive techniques of Position-aware, Orientation-aware and Waypoint-aware are designed to understand users’ movements and their surrounding geographical environment to solve the mentioned errors. The proposed Position-aware service samples received GPS data into motion states by Newton’s Law of Motion and derived Newton Markov Model (NMM). Based on NMM, Position-aware service builds up a stable Transition Probability Matrix (TPM) and learns the carrier’s behavior in the training phase. From the perceptive data of the stable TPM and online received GPS data, Position-aware infers the behavior of the GPS carrier to verify the rationality of the GPS data and then interactively rectifies received GPS data online. The Orientation-aware is done by cross reference GPS course data and digital compass. If users don’t follow the planning path, the Waypoint-aware service exercises to infer whether users are lost or not to shorten the disorientation period under the navigation state transition. Finally, the augmented navigation service is further integrated to provide an intuitive navigation interface to solve human cognitive errors. Through the interaction of these services, Live-view GPS navigation system ensures that user’s cognitive map is consistent with the navigation map. This search not only actually implements proposed Live-view GPS navigation system, but also conduct a series of experiments o validate the perception and usability of proposed perceptive services. The user studies confirmed the proposed Live-view GPS navigation system would supply a much more timely, intuitive and convenient navigation services than ever legacy GPS navigation system did.
