智慧型交通號誌燈控制研究與分析

Abstract

在現今各大城市中，壅塞的交通一直是亟欲解決的問題，而隨著資訊產業的快速發展，智慧型交通運輸的研究漸趨熱門，其中交通號誌燈的控制是智慧型交通運輸中重要的一環。交通號誌燈控制的研究分為兩個種類，一種是固定時間式的交通號誌燈控制，也就是事先幫每個交通號誌燈安排好紅綠燈的循環週期，藉此達到疏通交通阻塞的目的。而另一種是變動時間式的交通號誌燈控制，藉由動態的評估車況來配給紅綠燈的時間，以應付隨時變化的交通狀況。

在本論文的研究中，我們選擇研究變動時間式的交通號誌控制，我們會設計不同的車流拓樸 (traffic pattern)、不同的時間間隔、以及不同的車流資訊蒐集程度，來深入的研究變動時間式的交通號誌控制演算法的運作情形，並觀察演算法在不同情境下的運作情形，然後提出改善的建議，透過模擬案例實驗證實，我們的改良確實有效。

In today's major cities, traffic congestion has been a severe problem. With the rapid development of information industry, intelligent transportation system (ITS) research is becoming more and more popular. Traffic signals control is getting well concerned among ITS. Traffic signals control's study is divided into two categories. One is fixed time-based traffic signals control, that is, arranges the red-green cycle for each traffic signal to solve the traffic congestion. The other is dynamic time-based traffic signals control, which allocates the traffic signals by evaluating traffic flows dynamically to cope with the varied traffic conditions.

In our thesis, we choose a dynamic time-based traffic signals control mechanism to study its behaviors and characteristics under various kind of traffic patterns, time intervals, and collection of car information to get well understanding on the mechanism. We also propose our improvements on this algorithm and verify our improvements using NCTUns network simulator. The results show that our improvements perform better than the original algorithm.