請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/6196
標題: | 混合車流等候結構之號誌分流最佳化模式 A Study of Mixed Traffic Queuing Pattern on Separed-Flow Signal Optimization Model |
作者: | Hsin-Hung Liu 劉信宏 |
指導教授: | 許添本(Tien-Pen Hsu) |
關鍵字: | 混合車流,分流設計,等候型態,號誌最佳化模式,混合整數線性規劃, Mixed traffic,Separed-flow design,Queuing pattern,Signal timing optimization,Mixed-Integer-Linear-Programming, |
出版年 : | 2013 |
學位: | 碩士 |
摘要: | 本研究考量混合車流路段推進與路口停等之等候結構行為,提出一套號誌分流最佳化方法。過去研究處理混合車流的問題,多數將不同車種以小客車當量標準計算號誌時制。此方法在穩定且同質的車流環境下具有一定分析能力,但無法呈現不同車種駕駛行為所造成的巨觀車流差異。為了克服此方法之缺點與改善混合車流號誌時制演算之準確性,本研究發展一套時間與空間分流之號誌最佳化方法,考量市區車流等候型態特性加以模式化,描述混合車流交互作用與號誌影響關係。本模式結合路口幾何設計、車流變化與號誌時制三大構面成為整合式號誌最佳化程序,藉由實際車流資料進行統計檢定,發現車流等候型態與車輛數、到達順序有關,相同型態可歸納為同一車道群組,不同路段等候型態亦存在差異。
本文最佳化模式分為兩階段,先針對混合車流道路幾何進行最佳化,再根據最佳的空間配置與車流變化調整號誌時制。現實社會中號誌的重整大多著重於時制的微調,而二階段的設計方法不僅使道路與車道空間利用更具彈性,也避免因車流變化導致道路工程頻繁的變遷。第一階段模式經由敏感度分析發現車道與機車待轉區配置隨著流量與車道數不同而有變化,無論對稱或非對稱路口皆適用。第二階段定義之混合車流等候型態可以表現不同汽車與機車停等與紓解的組合,並將差異反映在號誌最佳化結果中。 本研究以假設之涵蓋4個路口的路網進行測試,並將最佳化結果與號誌最佳化軟體TRANSYT-7F比較,並以微觀車流模擬軟體VISSIM進行績效評估。在不同流量情 境下,低流量與中流量模式表現與TRANSYT-7F相似,部分路口汽車與機車平均延滯略有改善,高流量情況下則模式效果較不穩定。經由等候型態分析結果,本模式能呈現路口不同車道群等候線消長情形,對於混合車流號誌設計提供一套新的整合最佳化方法。 This research focuses on signal control modeling under mixed flow propagation and queuing patterns. In the past, many studies treated different modes of vehicles as the same standards and converted them into passenger car units. Although such conversion may be valid for stable and homogeneous traffic flow, it can’t reveal driving characteristics for different modes. To overcome such disadvantage and improve efficiency of mixed flow signal timing, this research develops a separed-flow signal optimization model with queuing patterns to describe the relationship between mixed traffic and signal timing. This model integrates geometric design, traffic variations and signal timing into its optimization procedures. According to queuing pattern statistical ratio test, we find out lanes with the same pattern can be classified as one group, but it related to number of mixed-flow vehicles and arriving sequences. The patterns differ from different road sections. The model divided into two stages: The first stage is lane-based optimization method with mixed flow; the second stage adjusts signal timing plan through proposed queuing patterns based on mixed flow area propagation and discharge under first stage design. In reality, the traffic signal re-timing programs mostly focus on fine tuning current signal design. The proposed two stage model provide not only more flexible way in lane space use, but avoid traffic flow variation resulting in frequent traffic engineering works. Through sensitivity analysis, this research reveals the lane marking assignment will vary with traffic volume and maximum number of lanes, and this model is applicable in either symmetrical or asymmetrical junctions . The queuing patterns can represent different car and scooter queuing and discharging combinations and reflected in signal optimization model. This research uses a test network with four intersections to study, and the result of proposed method will be compared to TRANSYT-7F. These optimal signal timing plans will be evaluated with VISSIM simulation tool and analyze its network performance. Through different flow scenarios, the performance of proposed model is similar to TRANSYT-7F in low or medium traffic demand, but it reduces when traffic volume is high. According to the results, this model considers the variation of mixed-flow queue in different lane groups, providing a new method in mixed-flow signal timing optimization. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/6196 |
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-102-1.pdf | 13.62 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。