請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/92160
標題: | 考量車聯網環境下停等車隊長度之即時連鎖優先號誌控制 A Real-time Coordinated Transit Signal Priority Control Based on Queue Length Prediction in a V2X Environment |
作者: | 黃信凱 Hsin-Kai Huang |
指導教授: | 許聿廷 Yu-Ting Hsu |
關鍵字: | 優先號誌,號誌連鎖,車聯網,停等車隊長度預測,動態規劃, Transit Signal Priority,Signal Coordination,V2X,Queue Length Prediction,Dynamic Programming, |
出版年 : | 2024 |
學位: | 碩士 |
摘要: | 為減少輕軌列車於號誌路口的停等延滯,並提升整體系統營運服務水準,輕軌優先號誌之設計和執行有其必要性。在臺灣現行的輕軌系統中,優先號誌的偵測機制以定點為主,然而,當衝突方向的車流量增加時,執行優先號誌將使一般車輛於路口的延滯明顯增加,對道路交通帶來負面影響。此外,僅考慮單一路口而未考量號誌連鎖易使輕軌列車仍須於下游路口停等,導致輕軌總旅行時間並無顯著下降。總結以上,問題主要源於無法即時獲取輕軌列車及一般車輛之位置與速度資料。近年,隨著車聯網環境的建置與測試,得以蒐集車輛位置與速度資料。然而過往研究多將車聯網技術應用於單一路口的即時號誌控制最佳化,部分研究則考慮上下游路口間的有效優先號誌連鎖,但少有研究針對兩者提出整合之探討。
因此,基於車聯網技術,本研究旨於建立一考量多路口連鎖之即時優先號誌控制邏輯。在輕軌列車觸發優先號誌後,將下游路口車輛依照方向、抵達時間分為三個類別,並透過停等車隊長度預測模型計算執行優先號誌後所造成的車輛額外延滯。從而利用動態規劃演算法,推估執行優先號誌在輕軌沿線各路口對於列車及一般車輛造成的停等延滯。 本研究透過車流模擬針對淡海輕軌系統進行案例分析,選擇其中五個連續路口建立模擬測試環境。結果顯示,在執行輕軌優先號誌後,單一路口輕軌延滯減少近8秒,而對於一般車輛僅增加1秒的平均路口延滯。在特定的時差及車流組合下,相較於僅考慮單一路口的優先號誌,考量號誌連鎖能減少約68.9%的輕軌平均路口延滯。此外,敏感度分析結果顯示,在衝突向每車道車流量大於400 (pcu/hr) 時,使用絕對優先策略將導致衝突向延滯上升147%,而納入車聯網的優先號誌策略可使之僅上升10%,有效降低輕軌優先號誌所帶來的負面影響。 本研究提出的優先號誌邏輯能同時考慮輕軌的總營運時間以及沿線各路口一般車輛的延滯。對於輕軌營運單位以及交通管理機關而言,在車聯網環境建置完成後,透過本研究提出的邏輯,能根據車流及輕軌營運特性,選擇適切的優先號誌方案。 In light of the delay of Light Rail Transit (LRT) at signalized intersections and the potential of degraded service levels, it is necessary to implement Transit Signal Priority (TSP) to enhance LRT operation. Among the current LRT systems in Taiwan, the vehicle detectors for TSP are primarily deployed over fixed points. However, upon high traffic volumes, TSP may significantly increase the delay for non-priority vehicles. Furthermore, designing TSP for individual intersections without considering the signal coordination may still cause waiting time for LRT vehicles at downstream intersections, resulting in limited benefits in reducing the tram travel time. Overall, the main challenge arises from the inability to obtain real-time locations and speeds of both LRT and general vehicles. In recent years, V2X (Vehicle-to-Everything) communications have been built and tested to collect vehicles’ locations and speeds. Previous studies have mostly focused on the applications of V2X technology to single intersections, while some researchers have sought to design effective TSP coordination across upstream and downstream intersections. However, relatively few research efforts have been made to integrate these two aspects. Accordingly, this study aims to develop a real-time TSP control logic in a V2X environment. Once the TSP mechanism is activated by a tram, general vehicles at downstream intersections are categorized into three types based on their direction and arrival time at the intersections. Based on a queue length prediction model, the additional delay caused by the TSP control is calculated. Subsequently, by adopting a dynamic programming algorithm, the waiting time of LRT vehicles and the overall delays at each intersection along the LRT route are estimated. The case study is made based on Danhai LRT system, where five consecutive intersections are selected to construct the micro-simulation platform for testing the capability of the proposed method. The experimental results show that the suggested TSP can lead to an 8-second decrease in the delay of trams at each intersection and averagely only a 1-second increase in the delays of general vehicles. By setting specific offsets for certain traffic flow conditions and considering signal coordination across intersections can even reduce the delays of trams at each intersection by 68.9%. Moreover, sensitivity analysis results indicate that when non-priority traffic flows exceed 400 pcu/hr per lane, implementing unconditional TSP strategy increases the delays of general vehicles up to 147%, while V2X-based TSP only results in a 10% increase, effectively mitigating the negative impact. The proposed TSP logic in this study takes into account both tram travel time and the delays of general vehicles at each intersection along the LRT route. For LRT operators and relevant government authorities, once the V2X environment is built, the proposed logic in this study can be utilized to select appropriate TSP strategies with respect to the characteristics of LRT operations and local traffic conditions. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/92160 |
DOI: | 10.6342/NTU202304191 |
全文授權: | 同意授權(限校園內公開) |
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-112-1.pdf 目前未授權公開取用 | 3.28 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。