應用車隊擴散理論於構建網路型適應性號誌控制模式之研究

• Main Author: 蔣封文
• Other Authors: 何志宏
• Format: Others
• Language: zh-TW
• Published: 2000
• Online Access: http://ndltd.ncl.edu.tw/handle/93200419429838179412

碩士 === 國立成功大學 === 交通管理學系 === 88 === In the recent years, urban traffic problem is always one of the important issues for many countries in the world. Among various kinds of traffic manage and control strategies nowadays, a brand new one, i.e. the traffic-adaptive signal control strategy, has been developed and expressed excellent performance in dealing with those urban traffic problems. However, this control strategy got used to employ microscopic traffic simulation model to handle its roadway vehicle operation in the past. So some problems could usually be occurred in their signal network building process. Besides, since microscopic traffic simulation models adopt the so-called car-following model to deal with roadway vehicle movement, the inherited characteristics of platoon dispersion has been always ignored. For these reason, this thesis attempts to integrate a self-developed traffic-adaptive signal control strategy with another macroscopic traffic simulation model, so as to become flexible and expendable while in application of it. Furthermore, the suitability for applying this concept into our traffic-adaptive signal control strategy could also be surveyed in the study. According to the testing outcomes, we find that our traffic-adaptive signal control strategy presents better performance than traffic-actuated as well as pre-timed control strategy in terms of average vehicle delay times, although it is slightly inferior to them in terms of percentage of stops. While integrating these two performance indices, still we find our traffic-adaptive signal control strategy has significantly the best performance. Meanwhile, we find that to adopt the horizontal queue back-up function will be more accurate than adopt the vertical queue back-up function, and vehicle delay time will be reasonably increased accordingly. On the other hand, we find that the computer operation time is shorter in executing the traffic prediction function by using the macroscopic traffic simulation model than using microscopic traffic simulation model. Based on these results we are convinced that the ideas and concepts employed in this study will have considerable possibility to be developed in the near future.