標題: 即時資訊狀況下駕駛人路線選擇行為之分析
The Analysis of Route Choice Behavior with Real Time Information
作者: 黃燦煌
Tsan-Huang Huang
陳武正
馮正民
Wu-Cheng Chen
Cheng-Min Feng
運輸與物流管理學系
關鍵字: 模擬系統;路線選擇行為;即時資訊;simulation system;route choice behavior;eal time information
公開日期: 1999
摘要: 目前國內有關智慧型運輸系統之發展計畫,多屬零星而片段,更缺乏一可供測試與評估智慧型運輸系統績效之模擬環境,以作為評選智慧型運輸系統最適發展方案之平台。就經濟、社會層面而言,智慧型運輸系統相關方案多須投入相當之人力、技術及經費方可運作,而其系統績效所影響之範圍又屬多元化且全面性,若未經方案測試與評估而冒然實施,必造成龐大之社會成本;就學術研究而言,與智慧型運輸系統相關之學理或模式必須經由測試及驗證其可行性,而目前可供測試用之真實交通資料零星且不易取得,藉由模擬程式模擬交通狀況之運轉,作為蒐集測試資料之途徑,以供智慧型運輸系統相關模式研發與改良之用,為當前較為可行之方式。 本研究即針對模擬系統中頗為重要的駕駛人選擇行為進行探討,且主要著重即時資訊狀況下駕駛人選擇行為分析,透過問卷調查及日誌式調查方式蒐集台北市駕駛人路線選擇的特性及可能影響路線或出發時間變動的因素,並針對即時交通資訊狀況下駕駛人路線選擇行為分析探討可行的方法及建立完整的分析架構。本研究以即時資訊符合度及相關性作為衡量資訊提供績效之依據,此種期望接受相關資訊引導的符合程度可分為資訊相關率、資訊錯誤率、資訊正確率及資訊誤判率,並分別建立其關係式及即時交通資訊影響模式架構。 依照模擬結果,當資訊相關率為0.9,資訊錯誤率為0.1,資訊正確率為0.9及資訊誤判率為0.1的時候,羅吉特路線選擇模式之績效明顯提昇,平均旅行時間可節省約12.83%~22.57%。本研究以平均旅行時間來看,在一般即時交通資訊中隨著路網中交通量的增加並沒有明顯依照資訊的詳細程度,而有平均旅行時間節省的差異,主要可能因為都市路網中駕駛人對所需選擇路線的狀況多能稍為瞭解;但是當道路封閉施工或突然發生事故時,由於駕駛人所面對的交通情況較無法掌握,因此詳細的即時資訊對路網績效較有幫助。 對於不同的可變資訊標誌設置比例對路網績效的影響,可以發現一即有趣的現象,即當路網中交通量較少時,不同的可變資訊標誌設置比例對路網績效的影響差距約在平均旅行時間5分鐘的範圍,但是當交通量逐漸增加時,對路網績效的影響差距增加至約在平均旅行時間10分鐘的範圍,隨著交通量增加約為原來交通量3倍時,不同的可變資訊標誌設置比例對路網績效的影響差距又逐漸縮小,約至原來交通量5倍時,其又回到原來差距之規模。 至於在駕駛人選擇行為逐日動態變化方面,駕駛人有以每一週的旅次經驗學習的型態,經參數校估結果,其中出發時間動態變換模式主要以動態基準為佳,其中又以動態基準10及30分鐘之模式校估結果較佳;至於行駛路線動態模式則以常用路線基準模式為佳。經模擬結果,大致上其平均旅行時間呈現下降的趨勢,且約至第三週以後(約為第十二天以後)其減少的幅度加大,顯然駕駛人已能透過學習的過程,而減少平均旅行時間。
Projects with regard to ITS are being developed in Taiwan now, but we are in lack of one simulation environment in which we can test or evaluate the performance of ITS. Because the projects of ITS usually requires too much labor, technique, budget, etc., it's important to develop a platform of ITS which could save much social cost for a lot of demonstrative projects. In the research area, models or theorem of ITS sometimes need to be tested or calibrated, too, but the traffic data in the real wold are usually difficult to collect for the use of calibration of ITS. The study used a two-step survey with the revealed and stated preferential parameters. The experiment recorded the drivers' pre-trip route choice behavior at 12 zones of Taipei. A logic model formulation was also adopted to form sequential route choice behavior. The results indicate that drivers update their knowledge of the system on a weekly circular basis and take routine driving routes in some range of routine departure hours. The calibration of departure hour concordance model indicates the two best alternations are dynamic 10 minutes and 30 minutes. The calibration of route choice concordance model indicates the best alternation is the routine route. In the simulation output, when the information related index equal 0.9, information error index equal 0.1, information correct index equal 0.9, and information miss index equal 0.1 could have good promotion for the Logic Model. Average travel time could save 12.83%~22.57%. During the traffic volume increasing, the detail of information didn't have clear difference in the travel time saving of the urban network. The range of travel time saving for the various amount of VMS in the network, could change from 5~10 min with the traffic volume up and down. The results indicate that drivers update their knowledge of the system on a weekly circular basis and the more travel time saving on the third week for the simulation model.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT880118039
http://hdl.handle.net/11536/65275
顯示於類別:畢業論文