請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/5227
標題: | 高速公路可變速限聯合匝道儀控最佳化模式 An Optimization Model of Coordinated VSL and Ramp Metering Control on Freeway |
作者: | Chun-Kai Chang 張鈞凱 |
指導教授: | 許添本(Tien-Pen Hsu) |
關鍵字: | 可變速限,匝道儀控,數學規劃,適應性控制, Variable Speed Limit,Ramp Metering,Mathematical Program,Traffic-Responsive Control, |
出版年 : | 2014 |
學位: | 碩士 |
摘要: | 隨著私人車輛持有率和使用率不斷增加,造成交通量迅速成長,在尖峰時段或連續假日高速公路往往會發生大範圍及長時間的壅塞,甚至回堵至地方道路而造成全面性的壅塞。這些壅塞將導致高速公路運作的效率降低,也帶來了龐大的社會成本。因此,為提升高速公路的使用效率以及減少社會成本,有必要採取相關的交通控制措施以維持車流的穩定。
本研究以國道五號北上路段為例,根據車輛偵測器資料分析研究範圍內之壅塞特性,由分析結果可知,國道五號北上路段壅塞範圍大多集中於雪山隧道,且車流速率於五分鐘內即會由90km/hr降至10km/hr,具有劇變現象。因此,本研究透過數學規劃法建構一套可變速限聯合匝道儀控最佳化模式,希望利用可變速限之特性控制車流以漸變的方式進入雪山隧道,並搭配上游匝道連鎖控制,以減緩或避免雪山隧道壅塞產生。研究中並以離散型二階巨觀車流模式-METANET為基礎,發展一套適用於國道五號北上路段之巨觀車流模式,本研究將之稱為「METANET長隧道模式」。 然而,最佳化模式因參數校估不易以及求解速率過慢等特性,使之難以運用在線上控制中,為克服上述問題,本研究利用類神經網路建構預測模式,以預測最佳化模式之輸入參數,並透過啟發式演算法提升最佳化模式之求解效率,以達到線上控制之目的。 為評估模式之控制績效,本研究利用微觀模擬軟體VISSIM繪製國道五號北上路段之模擬路網,並透過驗證程序確認模擬路網之真實性,最後,以VISSIM外掛程式VisVap將最佳化模式求解結果應用至模擬路網中,模擬結果顯示,在控制後主線旅行時間與雪山隧道總通過量均有顯著的改善,顯示本研究所研擬之控制模式能有效提升國道五號北上路段之使用效率。 With the growth of private car holding rate and usage rate, traffic volume increases rapidly, huge scope and long-time congestion happens in the peak hour or on long vacations on freeway, the congestion also blocked back to local roads which cause entirely congestion. The efficient of freeway reduced and social cost increased due to congestion, in order to raise the efficient and reduce social cost, it is necessary to adopt relative traffic control strategy to maintain the stability of traffic flow. The research takes north bound of freeway No. 5 as an example. First of all, to analysis the congestion characteristics in the research scope by the data from vehicle detector. It is known that the congestion segment is focused on HsuehShan tunnel, and car velocity reduced from 90km/hr to 10km/hr in 5 minutes, which is the catastrophe phenomena. As a result, this research construct a optimal model of coordinated VSL(Variable Speed Limit) and ramp metering by mathematical program, to control traffic flow entering HsuehShan tunnel with gradually speed reduction by the characteristic of VSL, also coordinate with upstream ramp interlocked control, to mitigate or even prevent from the congestion of HsuehShan tunnel. This model also based on discrete second order macroscope traffic flow model - METANET to develop a macroscope traffic flow model witch fits the north bound segment of freeway No. 5, which is named 'METANET long tunnel model' by this research. However, due to the difficulty of parameter calibration and slow solving rate of optimal model, it is unrealistic to be used in online control. In order to overcome this difficulty, Artificial neural network is used to construct the prediction model and predict the input parameters of optimal model, also heuristic algorithm is used to enhance the solving rate, with the combination to reach the goal of online control. In order to assess the performance of this model, microscope simulation software - VISSIM is used to sketch the network of the northbound of freeway No. 5, the validation process is taken to confirm the reality of simulation network. Finally, the plug-in software VisVap is been used to apply the result optimal model into simulation model. The result indicated that the mainline travel time and total throughput of HsuehShan tunnel increased significantly after control, which indicate the usage rate of northbound of freeway No. 5 is enhanced effectively by the control model in this research. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/5227 |
全文授權: | 同意授權(全球公開) |
顯示於系所單位: | 土木工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-103-1.pdf | 9.82 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。