交通號誌配時調整對路口延滯最小化之探討：考量行人優先

Abstract

隨著都市化發展，市區道路在尖峰期間難免發生壅塞，且行人與車輛的互動甚或衝突頻率日益增加。另一方面，隨著人本交通環境意識的提升，近年來臺灣各地方政府積極致力於改善交通環境；法規方面亦將駕駛未禮讓行人的罰鍰提高，期望提升行人安全。然而，在存在右轉車輛與穿越行人同綠的路口設計下，考量行人優先原則，右轉車輛需禮讓行人，導致後方車隊增長，平均延滯時間上升。因此在保障行人安全的同時，如何兼顧交通效率，避免延滯過度增加，亦是甚須探討的問題。 本研究著重在右轉交通車流效率改善，以行人優先為前提，檢視右轉車輛的飽和流率受行人阻擋影響的變化，透過調整行人綠燈長度，以達成兩項最佳化目標：一是最小化行人與右轉車輛的整體平均延滯，二是最小化兩者平均延滯的差異。不同於現行設計手冊僅以單一行人阻擋折減因子處理人車衝突，本研究進一步考量人車同綠情境下的實際紓解行為。行人常於綠燈初期大量通行，而右轉車輛則需待行人紓解率與抵達率趨於平衡後，才逐漸恢復應有的紓解能力。此種人車互動關係可劃分為 12 種型態，本研究並依此建立多流動延滯模型，用以呈現行人與車輛的抵達線與紓解線之交會平衡時間。模型假設行人與車輛抵達率為常數，行人紓解率受行人號誌變化影響，右轉車紓解率則同時受行人紓解變化與車輛號誌控制所影響，整體以分段線性方式表示。 本研究針對臺北市六個主要路口的行穿線與衝突右轉車道進行觀測，以人車衝突區域作為計數區，取得行人與車輛的抵達與不同時段的紓解數據。彙整資料後發現，穿越行人與右轉車輛的紓解率呈對數關係，此結果可作為後續最佳化模型的參數輸入。研究結果顯示，若人車抵達率處於一般狀況，建議給予最長的行人綠燈時間以獲得整體最小延滯；若考量人車平等，則行人綠燈長度通常會建議調整至可行解範圍的中間區段，以達到人車延滯最小差異之目標。

With ongoing urbanization, traffic congestion in urban areas has become inevitable during peak hours, and interactions or even conflicts between pedestrians and vehicles occur frequently. Meanwhile, as awareness of humanity-oriented traffic environments continues to grow, local governments in Taiwan have been actively working on improving road facilities, and for the regulatory aspect, penalties for drivers who fail to yield to pedestrians have been raised, all to pursue in enhancing pedestrian safety. However, in intersection designs where right-turning vehicles and crossing pedestrians share a concurrent green phase, prioritizing pedestrians generally means right-turning vehicles must yield, which can result in longer queues and increased average delays. While ensuring pedestrian safety, how to maintain traffic efficiency and prevent excessive delay is also worth further investigation. This study focuses on improving right-turn traffic efficiency under the premise of pedestrian priority. By adjusting the pedestrian green duration, the study proposes two optimization objectives: (1) minimizing the overall average delay of pedestrians and right-turning vehicles, and (2) minimizing the difference between the two delays. Unlike existing design manuals that address pedestrian blockage using a single adjustment factor, this study incorporates real-world discharge behavior under concurrent green conditions. Pedestrians tend to discharge in greater volumes during the early part of the green phase, while right-turning vehicles gradually recover their dischargeability as the pedestrian discharge rate reaches arrival rate. These pedestrian-vehicle interactions are categorized into 12 cases, forming a multi-movement delay model that identifies the timepoint when arrival and discharge curves intersect. The model assumes constant arrival rates for both pedestrians and vehicles. Pedestrian discharge rates vary with pedestrian signal changes, while right-turn vehicle discharge rates are affected by both pedestrian movement and vehicle signal phases. All discharge trends are represented using piecewise linear functions. Field data were collected from six major intersections in Taipei City, focusing on pedestrian crosswalks and their conflicting right-turn lanes. The pedestrian-vehicle conflict zone was used as the counting area to capture arrival and discharge volumes during different time intervals. Analysis shows that the discharge rates of pedestrians and right-turning vehicles follow a logarithmic relationship, which is then used as parameter input to the optimization model. Results indicate that under typical arrival conditions, the longest pedestrian green duration yields the lowest total delay. When emphasizing equity between pedestrians and vehicles, the pedestrian green time is generally recommended to be set near the middle of the feasible range to minimize the delay difference between the two user groups.