以延滯指標評估多種交通號誌控制邏輯之適用時機 —以獨立路口簡單二時相為例

Abstract

隨著社會經濟的快速發展，都市中的車流量也極速的增加。在道路容量無法提供大量的車流需求時，往往會造成交通壅塞問題。為解決交通壅塞問題最常見的手法有道路基礎設施的擴建及交通管理策略。然而當道路基礎設施無法擴建時，便只有針對交通管理著手，其中號誌控制邏輯是很重要的交通管理項目之一。相對於早期號誌控制邏輯採用的定時控制(Time Of Day)已無法因應交通車流的需求變化，鑑於定時控制邏輯將會浪費時間、燃料及金錢，對此國內外皆有許多學者投入改善定時控制所產生之問題，而現今針對此問題之研究，已發展出許多更符合需求的號誌控制邏輯，如半觸動控制、全觸動控制、運用卡曼濾波、基因演算法或類神經網路於號誌控制邏輯、二元決策號誌控制邏輯，以及程序性決策號誌控制邏輯等。 雖然上述各種號誌控制邏輯模式在國內外已有許多相關研究，但比較不同號誌控制邏輯之運作績效以建立各號誌控制邏輯在不同車流情境下之適用時機卻較為瑣碎稀少。因此本研究比較數種不同型式之常用交通號誌控制邏輯，包括定時控制、全觸動控制、運用卡曼濾波之動態計算號誌控制，以及改良OPAC(Optimization Policies Adaptive Control)。利用微觀車流模擬方式模擬四種獨立四叉路口，並同步模擬不同號誌控制邏輯，再以指數分配函數產生不同交通車流情境，並以微觀車流模擬方式累計車輛停等延滯，分析比較此四種號誌控制邏輯在不同情境下之車輛停等延滯時間，獲得在不同交通車流情境下較佳實施之號誌控制邏輯。研究結果將可作為選擇號誌控制邏輯之判斷標準，依此訂定客觀的號誌控制邏輯切換準則，對未來發展智慧型號誌系統將能更有效應付各種不同車流情境之變化。

With the rapid development of economy, vehicle flow has increased alarmingly in the metropolis. This usually leads to traffic jams when the road capacity fails to satisfy the traffic demand. General methods to solve this problem include extending road infrastructure and setting up strategies for traffic management. If the former fails, better traffic management has to be adopted, one important object of which is traffic signal logic. Since earlier traffic signal control logic which adopted time-of-day (TOD) scheduling/pre-time control cannot meet the need of traffic flow and variation, many traffic signal control logics such as half-actuated control, full-actuated control, Kalman Filter, Gene Arithmetic, and Quasi-nerve Network adopted for signal traffic control logic and dualistic traffic control logic, and program-based signal traffic control logic were developed. Although the traffic signal control logics mentioned above have been found in many studies, the criterion for selecting an optimal control-logic along traffic conditions through assessments in the same level regarding their performances and are few. Therefore, four types of usual traffic control logics are compared in this paper including time-of-day pre-time control, full-actuated control, dynamic -computing control with Kalman Filter, and adaptive control improved from the Optimization Policies for Adaptive Control. A micro-traffic flow simulation synchronously operated with multiple control logics is developed herein to manipulate a signalized intersection in various traffic patterns. On the simulation, the number of passing vehicles, stop vehicles, total waiting time, and total delay time in each period are calculated among these four types of control under different traffic flow contexts. Based on the results and comparisons, the criterion for selecting an optimal control-logic along traffic conditions in terms of traffic parameters is obtained. The resulted criterion can be taken as online traffic control execution to switching control logic in terms of real-time traffic conditions as well as a reference of developing an intelligent signal controller or system.