本研究探討相較於當前使用靜態圖資計算導航路徑,或僅提供部分壅塞情報的導航系統,一個以分散式移動感應器來取得道路系統中使用者位置、速度等即時路況資訊的導航系統,是否能夠以真正即時的路況資訊為考量,推薦出避開壅塞、減少通勤時間,更符合使用者期望的導航路徑。由於沒有實際執行的軟硬體環境,本研究以系統模擬的方式進行實驗,比較使用兩種不同導航方式移動的系統在通勤時間方面是否有差異。實驗結果顯示,考量了全體移動物件在道路網路上的位置、速度而得出之壅塞資訊的新導航方式,比只考量靜態圖資、推薦路程最短的大路的一般導航更能省時、更有效率。而更進一步,本研究提出學習使用者的行為模式的模型,將使用者行為的個體差異納入導航路徑的考量,學習使用者的偏好,讓導航的產出更能滿足使用者的習慣使用者偏好學習的功能與效果。
This thesis studies a new approach of real time navigation which acquires real time traffic flow information by treating every user in the road network as a sensor, and collecting the information like speed and location from every sensor. Then in this thesis, because of the lack of real world platform to run the test, a simulation experiment is established to see whether the assumption is correct. The result of the simulation experiment shows that with real time traffic congestion condition being known, the new navigation approach has recommended the better routing path for moving objects then basic navigation approach, with only distant and road classes being know. The last part of this thesis brings up a future vision with user-preference-learning mechanism which can provide better navigation recommendation to suit users’ behavior.