號誌路口車輛行進軌跡之模化與延滯時間之比較

Abstract

號誌化交叉路口匯集各種不同方向車流，在車流密度過大或號誌時制設計不佳時，經常會成為都市交通的瓶頸，其影響路網績效甚鉅。平均延滯是一衡量交叉路口運轉績效的重要指標。延滯在評估道路設施服務水準與號誌設計等方面深受重視，以往在路口延滯的相關研究中，大多是估計停等延滯，再藉轉換因素來推估臨進路段延滯，惟停等延滯在某些情況下會出現較大的誤差，間接也影響到臨進路段延滯。而車隊時間延滯在定義上可以涵蓋停等延滯與臨進路段延滯，本研究欲以實證法探討以車隊時間延滯直接估計路口的延滯，並在不同需求或交通控制情況下，利用車輛軌跡圖解法推算各種延滯值。

有鑑於一般文獻所使用的延滯模式，大多以停等延滯為基礎，而直接引用國外的延滯模式也常無法兼顧國內的交通型態，所以本研究先分析並探討導致車輛延滯之因素，作為建構延滯模式的基礎。再以調查資料分析車輛進出號誌交叉路口的情形，整理出車輛在交叉路口停等位置與離開速率的關係。並由調查之延滯資料來驗證模式之正確性，最後利用本研究所建構之模式直接算得延滯值，作為評估交叉路口服務水準與交通控制策略之參考。

建構的延滯模式配合使用JAVA程式語言所撰寫出的模擬程式，利用追蹤車輛軌跡的方式，用圖解法計算出停等延滯、路口臨進路段延滯和車隊時間延滯。由模擬過程發現模式所估計之平均車輛延滯值，包含車隊時間延滯和停等延滯，與實際調查之延滯值相近，說明了模式的可用性。從敏感度分析上，車隊時間延滯比停等延滯和臨進路段延滯來的大。本研究亦探討在特殊情況下下，如：路段長度過長、等減速度值極小以及臨進速率值極大等，使用車隊時間延滯值的缺點。

Signalized intersections are the locations that convert each approaching traffic stream into merging, diverging, and intersecting movements. Such locations will become the bottlenecks that may further degrade the traffic system due to dense demand and/or worse signal installation. Delays have been long term known as one of the most important indices in evaluating intersection control performance, for instance, the level of service and/or signal timing design. It is seen that most of the past research had widely applied the stopped delay as a basis to estimate some others such as approach delays. Unfortunately, this may sometimes lead to a higher estimation error. With the above consideration, this study is intended to explore and measure the other term, time-in-queue delay, for intersections using the graphic method supported by field data.

Because most of the delay models were developed based on stopped delays and the traffic characteristics may vary greatly between countries, directly citing the existing models, approaches, or specific research results for delays may greatly deviate the true value in our country. Thus, this study first analyzed some key elements of delay time with which the analytical delay model could be formulated. Secondly, the location of each vehicle entering and leaving the signal at each time slice during the study period was traced and recorded. The trajectories of all individual vehicles were then plotted to verify the model validity. Finally, the analytical model was directly used to estimate various delay values for evaluating service quality and traffic control alternatives at intersections.

Based on the analytical model, a simulation program was developed in JAVA in order to generate successive vehicle arrivals and record their corresponding trajectories when approaching the signal. With the data recorded, the stopped delay, approach delay, and time-in-queue delay to individual vehicles could be estimated separately. The simulation results showed that the three estimated delays were fairly close to those collected at sites. A series of sensitivity tests also showed that the time-in-queue delays are larger than the other two. In addition, this study also discussed the drawbacks of using time-in-queue delay when special conditions such as long distance travel approaches, extremely small deceleration rate, and high approaching speeds occur.