張堂賢2006-07-252018-07-092006-07-252018-07-092004http://ntur.lib.ntu.edu.tw//handle/246246/2853為達到先進運輸管理系統之運作，本研究提出一即時交通資訊擷取系統， 規劃以計程車車隊支援探測車系統之主體，結合路側偵測系統，以對於都市地 區路網系統進行廣泛而完整之交通資訊擷取，並研擬系統運作下之資料庫系統 設計。根據此一資料系統架構之下，考量資料項目之取得，本研究分別建構「路 段旅行時間預測」和「動態旅次OD 推估」之數學模式，並以廣義最小平方法 和推廣卡曼濾波器進行模式之演算，對於路網系統之車流狀態進行預測，藉以 支援相關動態交通控制、管理之決策。 在模式驗證部分，本研究係透過Paramics 軟體模擬一般化棋盤型路網系統 下之交通車流，對路段旅行時間預測模式之預測結果進行分析，研究中主要藉 由準確度、強健性和穩定性三個面向評估模式之預測能力。預測結果顯示，經 由模式校估之過程並對於來源資料在演算前進行相關處理，可得到優良之預測 結果表現。對於動態旅次OD 推估模式則是同樣根據模擬結果進行一試算流程， 將推估所得之旅次OD 流量反應於路段流量上可得到良好之預測結果，是以評估模式之推估結果為合理。 根據預測誤差，探討探測車回傳資料之於整體車流之代表性在模式預測準 確度之影響，分析結果顯示，在車流中探測車佔有率達到5%以上，本研究所提 出之路段旅行時間預測模式即可反應良好之預測能力。Abstract In order to achieve advanced traffic management system (ATMS), This study proposes a real-time traffic data acquisition system. The framework of system has its basis of taxi fleets as probe vehicle system, and combines roadside detectors to collect traffic data from urban network extensively. According to the physical architectural, the study builds the mathematic models of “travel time prediction for road section” and “dynamic OD estimation”. The algorithms are based on generalized least squares (GLS) and extended Kalman filter (EKF) repectively. Through the prediction model, the system state of traffic flow is predicted to support decision-making of traffic control and management. To verify the models, for travel time prediction for road section, this study analyzes the prediction results of grid network simulation from Paramics, and evaluates prediction ability by indices of precision, robustness and stability. It proves that well prediction results are obtained through calibration and data processing. For dynamic OD estimation, the study calculates traffic count prediction from estimated OD flows, which shows the mean error is within 10%. Therefore, it concludes the model is reasonable. Also, considering the represenstative of probe vehicles to traffic flow, the model has well prediction ability, when the penetration of probe vehicles is above 5%.application/pdf701499 bytesapplication/pdfzh-TW國立臺灣大學土木工程學系暨研究所交通資料擷取探測車路段平均旅行時間動態旅次ODTraffic Data AcquisitionsProbe VehicleTravel Time PredictionDynamic ODreporthttp://ntur.lib.ntu.edu.tw/bitstream/246246/2853/1/922211E002079.pdf