【作者】 杨海东；

【导师】 王万良； 杨旭华；

【作者基本信息】 浙江工业大学 ， 控制理论与控制工程， 2007， 硕士

【摘要】 随着智能交通系统的发展,城市交通信号控制已成为最重要的研究方向之一。由于城市交通的复杂性,采用传统的控制方法已无法有效地解决交通信号控制问题,本文采用人工智能控制的方法对城市交通信号进行研究,主要包括以下几方面内容:1.基于RBF神经网络建立了单交叉口自学习控制系统。该系统能够模拟交通警察指挥交通的思维过程,能够根据四相位交叉口各相位车辆的排队长度,对各个相位的绿信比和总的信号周期进行实时分配,并且随着交通状况的变化,可以对信号配时效果进行评价,根据车辆排队长度的变化,对信号作出调整,具有自学习功能。与传统的定时控制相比较,该系统能够更好地适应实际交通状况,提高交叉口的通行能力。2.提出优化相序和模糊神经网络控制相结合的方法对交通干线进行实时控制,并建立了协调控制模型。该模型由两层控制器组成:交叉口控制器(下层)和协调单元控制器(上层)。下层控制器又由两个控制器组成:相序优化器和绿灯延时控制器,相序优化器用于调整交叉口的相序,绿灯延时控制器负责各个相位的绿灯延时时间。上层控制器用于调整干线的信号周期和两交叉口间的相位差,使干线上行驶的车流尽可能不遇到红灯,并且使各交叉口车辆排队长度尽可能短。仿真表明:该模型比传统的定时控制更能有效地减小平均车辆延误,验证了算法的有效性。3.最后,对全文进行总结,并对进一步的研究提出一些展望。更多还原

【Abstract】 With the development of ITS (Intelligent Transportation Systems), urban traffic signal control has been growing into one of the most important research aspects. As the complexity of urban transportation, traditional methods are not able to settle the problem of signal control efficiently. This paper deals with urban signal control by means of intelligent control algorithm, and focuses on the following aspects:1. a self-learning signal control system based-on RBF Neural Network is established. This system can simulate the traffic police’s experience. According to the queue length in each intersection, the system can give out both the signal cycle and the split of each intersection. Furthermore, it can evaluate the effect of the control with the changing of the traffic, and then adjust the signal. Simulation results reveal that the system can much more perfectly control the actual traffic condition and improve the passing ability of the intersection.2. a kind of real time arterial signal control method based-on phase sequence optimization and Fuzzy Neural Networks is put forward in this paper, and a coordinated control module is established using two-stage controller: intersection controller (The lower course) and coordinated controller (The upper course). The lower course of the module consists of two controllers: phase sequencer and green-time delayer. The two controllers are used to optimize the phase sequence and adjust the green-time delay respectively. The upper course of the module is responsible for adjusting the signal cycle of the traffic trunk road and the offset between two intersections. The results of simulation prove that the proposed module has better performance than traditional fixed-time control.3. In the end, summarizing the whole work and pointing out some content which would. be researched in the future.更多还原