透過您的圖書館登入
IP:3.21.104.109
  • 學位論文

利用行動通訊資料推估用路人旅行路徑及速率

Estimation of Travel Path and Speed Based on Cellular Data

指導教授 : 邱裕鈞

摘要


利用即時交通資料預測旅行速率是先進旅行者資訊系統(Advanced Traveler Information System, ATIS)中不可或缺的,而目前台灣用來推估旅行速率大多是利用車輛偵測器(Vehicle Detector, VD)或是搭載全球定位系統之探偵車輛(GPS-based Vehicle Probe, GVP)所偵測到的資料來進行推估,但是以上的設備常會因為設置、維護以及數量問題而無法推估整體路網之資訊,往往只能推估高快速公路及幹道等旅行速率。   近年來由於智慧型手機的普及,手機通話及上網的使用越來越普遍,利用無線電信網路追蹤使用者位置的方式來產生即時交通資訊的研究也因應而生。但目前針對移動定位設備進行旅行者路徑與速率推估的研究大多是針對GPS的定位資料進行演算法研擬,但是行動通訊的定位資料之誤差比GPS定位誤差大了許多,由於誤差極大,因此這些演算法均無法直接提供手機訊號之探偵車(Cellular-based Vehicle Probe, CVP)使用,而現今臺灣針對CVP的相關研究也僅能適用於國道等幹道路網。   因此本研究參考國外CVP相關研究,利用行動通訊資料之特性並結合拓樸地圖匹配演算法及K條最短路徑演算法,研擬一套利用行動通訊資料推估旅行路徑及時間的演算流程,並透過微觀車流模擬軟體產生4種交通狀況下之行動通訊及交通車流資料,據以進行驗證。   最後本研究將推估之路徑與實際路徑進行正確度分析,在推估路徑正確度的部分,本研究在四個時段的推估結果中,平均正確率高達92.72%,得到之結果相當良好。在推估速率正確度方面,本研究比較40%、30%及20%之不同手機涵蓋率下所推估之路段旅行速率正確度,在比較結果可以看出當CVP涵蓋率越低時,其推估速率的誤差率也會提高,但在20%的涵蓋率下,其誤差率也僅有6.63%,尚屬合理誤差範圍,因此將此推估結果拿來當作下一時段之歷史旅行時間使用時並不會造成路徑推估上的影響,顯示本演算法之可應用性。本研究之研究成果可以提供相關產、官、學單位未來進行CVP實作與推估旅行資訊之參考。

並列摘要


Real-time traffic information is essential to the advanced traveler information system. The traditional devices to collect traffic information, such as vehicle detectors, GPS-based vehicle probe (GVP) are rather cost-intensive for fully spatiotemporal coverage. With the rapid popularity of cellular devices, use of cellular-based vehicle probe (CVP) to obtain travel information of a large-scale road network becomes viable. Most of similar studies using vehicle positioning techniques to determine travel time information are based on GPS technique which is much more accurate than CVP. Those GVP algorithms may not be suitable for CVP. A few proposed CVP algorithms are however only suitable for freeway systems.   Based on this, this study aims to propose a CVP algorithm based on the topological map-matching algorithm and K-shortest path to determine the travel path and time of travelers. To investigate the applicability and accuracy of the proposed algorithm, a case study on simulated cellular and traffic data under four traffic conditions is conducted.   The results show that the accuracy of predicted travel path reaching 92.72%, suggesting good performance of the proposed algorithm. As to the travel time prediction, under three cellular penetration rates of 40%, 30% and 20%, the performance of the travel time prediction slightly deteriorates as the penetration rate decreases, but even for the lowest penetration rate of 20%, the algorithm is still able to maintain its error rate as low as 6.63%, suggesting the applicability of the proposed algorithm.

參考文獻


鍾志成, 林蓁, 林志偉, 林誌銘, 2012. 利用旅客流模擬模式計算捷運系統人流服務指標之研究. 運輸計劃季刊, 41(3), 279-310.
Sun, D. H., Wang, C. L., 2007. Map matching algorithm based on fuzzy pattern recognition for vehicle location. Computer Engineering and Applications 42(25), 227-230.
Bar-Gera, H., 2007. Evaluation of a cellular phone-based system for measurements of traffic speeds and travel times: A case study from Israel. Transportation Research Part C 15(6), 380-391.
Chaturvedi, M., Srivastava, S., 2013. Real time vehicular traffic estimation using cellular infrastructure. Advanced Networks and Telecommuncations Systems (ANTS), 2013 IEEE International Conference, 1-6.
de Queiros Vieira Martins, E., 1984. An algorithm for ranking paths that may contain cycles. European Journal of Operational Research 18(1), 123-130.

延伸閱讀