基于车流量的交通灯动态调整策略的研究与设计

【作者】 黄向党；

【导师】 罗瑜； 杨雄；

【作者基本信息】 电子科技大学 ， 软件工程（专业学位）， 2012， 硕士

【摘要】 交通灯是管理城市交通的重要工具,交通灯对道路交通流的影响近年来引起广大学者的广泛关注。目前绝大部分的交通灯时间都是预先设定好的,所谓静态,或者非动态配时。不管是车流高峰还是低谷,红绿灯的时间及间隔都固定不变,显然地,在这样的配时策略导向下,车流辆多时不能合理发挥交通导向作用,车流辆少时不但不能提高车辆的通行效率反而造成不必要的延误。针对这一问题,当前信号灯动态配时策略研究主要采用的算法有模糊理论、神经网络、和遗传算法三大类。本文提出一个基于遗传算法的动态配时策略,主要有以下两个创新点：1.针对目前单一指标研究的不足,本文采用多指标控制策略。例如,在车流密度,延误时间,平均车速等多个不同交通流量指标的基础上展开研究。2.针对目前单一交叉路口指标控制研究的不足,本文综合整条线路上多个交叉路口的交通流量指标,提出一个基于遗传算法的多目标优化组合的交通灯配时策略,仿真结果证明了该策略的有效性。更多还原

【Abstract】 The traffic lights is an important tool for the management of a city, in recent years, the impact of traffic lights on the road traffic flow has aroused widespread concern in the majority of scholars. Most of the current traffic signal time is set in advance, this is the static or non-dynamic timing plan, whether the traffic peak or trough, red and green light time and intervals are fixed, obviously, preset signal control strategy is not applicable in the case of traffic peak, on the contrary, it would not only decrease the traffic efficiency of vehicles but also increase the vehicles average delay in that case of trough. To solve this problem, the research on the signal timing plans mainly adopts three algorithms, including fuzzy theory, neural network and the genetic algorithm. This thesis proposes a dynamic timing plan based on the genetic algorithm, which mainly in the following two innovation points:1. In view of the current research deficiency of single indicators, this thesis adopts Multiple Indicators control strategy, research on different traffic flow indicators, including traffic flow density, delay, vehicle average speed etc.2. In view of the current control research deficiency of a single crossroad indicator, in this thesis, based on the comprehensive consideration of traffic flow indicators on multi-crossroad of a whole line, a multi-objective optimization traffic light timing plan based on genetic algorithm is presented. At last, the simulation results verify the efficiency effectiveness of this plan.更多还原