號誌化路口車流衝突量模式之研究

A Study On Traffic Flow Conflict Models for Signalized Intersections

號誌化路口往往為都市地區道路系統之主要交通瓶頸，因此如何有效地針對路口交通特性加以分析，並適時提供一具有安全、效率及順暢的控制措施或策略，則是研究號誌化交叉路口的重要課題。以往對於交叉路口安全性之評估，大多偏重於衝吏，點之計算及肇事頻率多寡之分析，然而衝突，點數目僅能解釋路口幾何設計與號誌時制計畫對路口安全性之影響，而難以反映路口交通量與安全性之關連性;至於肇事頻率則僅能解釋已發生之衝突，亦無法反映衝竟發生之潛在機率。因此，本研究依據號誌化路口之各種時相設計特性，利用衝突量模式之理論基礎加以推演，並分析潛在機率、流量大小及車輛位置等因素對號誌化路口衝突之影響，據以建立號誌化路口之車流衝突量模式。研究結果顯示，號誌化路口之車流衝突量模式，除可更真實反映不同號誌時相設計下，交叉路口整體車流之交叉、分出及併入等衝突行為外，更可增加交叉路口期望衝突量理論之應用範園，對於號誌化路口之安全性分析提供了良好的量化性指標，並可作為未來交通工程師設計與評估交通工程設施之參考依據。

Signalized intersections are always the main traffic bottlenecks of road system in urban areas. Therefore, how to analyze the traffic characteristics effectively and provide safe, efficient and fluency strategy in time are very important issues for signalized intersections. In the past studies, the evaluations of intersections' safety were mainly based on the analysis of conjlict points and frequency of accident. However, the confilct points analysis only explains the effect of geometry design and signal phasing design at intersections. The relationship between traffic flow and safety can not be shown. The frequency of accident analysis only explains confilcts that have happened and cannot be able to infer the potential probability of accident occurring. This research infers the theory of expected conflict model based on various phasing design characteristic of signalized intersections, analyzes the factors of potential probability, traffic volume and location of vehicles for conflicts occurring and effects. Base on the inference and analysis, the traffic flow conflict models for a signalized intersection are developed. This research shows that these models can reflect conflict maneuvers such as crossing, diverging and merging of vehicles under different signal phasing design, increase the application of the expected conflict model. It can also serves as a quantity criterion on intersection safety and reference for traffic engineer on designing and evaluating traffic devices of intersections in the future.