单点信号交叉口智能控制的优化模型和方法研究

发布时间：2018-10-04 19:31

【摘要】：随着城市化的推进和机动车保有量的不断增长,城市道路交通面临着巨大压力,拥堵问题层出不穷,而延误主要发生在交叉口,因此,交叉口的信号控制对道路交通的健康运行有着举足轻重的作用。由于传统的定时信号控制方法不能适应随机变化的交通流,可以根据交通流的变化实时的调整配时方案的智能控制方法逐渐成为提高交叉口通行效率的有效解决方案。首先,本文在传统信号配时的基础上提出了定时控制的优化模型。综合考虑延误、停车率和通行能力三个指标,建立了基于粒子群算法的多目标优化模型。在降低延误和停车率的同时尽量增大交叉口的通行能力,并通过实际案例分析,验证了模型的有效性。其次,本文研究了基于模糊控制的智能控制方法。模糊控制方法包含绿灯延时模块和相序优化模块,绿灯延时模块根据车辆排队长度调节绿灯时长,相序优化模块根据不同相位的对通行权的需求度来调整相位顺序。模糊控制方案可以根据不断变化的交通流智能地调整配时方案,降低交叉口的车辆平均延误。然后,在模糊控制的基础上,本文引入神经网络建立了模糊神经网络信号控制方法。该方法可以发挥神经网络的自主学习优势,利用大量的实际数据对模糊神经网络进行训练和学习,得到可以适应不同状况交通流的智能控制方案,有效提高交叉口的运行效率。最后,论文基于实际交叉口建立了交通信号控制仿真模型,对论文提出的三种控制方法进行了仿真分析。延误对比表明,模糊神经网络控制的控制效果最好,延误相对定时控制降低20%-30%,模糊控制次之,相对定时控制降低10%-15%。最后对模糊神经网络控制仿真过程中的车辆排队长度与配时方案的变化规律做了详细分析,讨论了排队长度与配时结果的关系,验证了控制方法的科学性。
[Abstract]:With the promotion of urbanization and the increasing number of motor vehicles, urban road traffic is facing great pressure, congestion problems emerge in endlessly, and delays mainly occur at intersections. The signal control of intersections plays an important role in the healthy operation of road traffic. Because the traditional timing signal control method can not adapt to the random change of traffic flow, the intelligent control method which can adjust the timing scheme in real time according to the change of traffic flow has gradually become an effective solution to improve the traffic efficiency of intersection. Firstly, based on the traditional signal timing, an optimal timing control model is proposed. A multi-objective optimization model based on particle swarm optimization (PSO) was established by considering delay, parking rate and capacity. At the same time, the delay and parking rate are reduced and the capacity of intersection is increased. The validity of the model is verified by practical case study. Secondly, the intelligent control method based on fuzzy control is studied. The fuzzy control method includes green time delay module and phase sequence optimization module. The green light delay module adjusts the green light time according to the vehicle queue length, and the phase sequence optimization module adjusts the phase order according to the demand of different phases for traffic weight. The fuzzy control scheme can intelligently adjust the timing scheme according to the changing traffic flow and reduce the average vehicle delay at the intersection. Then, on the basis of fuzzy control, this paper introduces neural network to establish a fuzzy neural network signal control method. This method can take advantage of the autonomous learning of neural network, train and learn the fuzzy neural network with a large amount of practical data, and obtain an intelligent control scheme which can adapt to traffic flow in different conditions, and effectively improve the running efficiency of intersection. Finally, the traffic signal control simulation model is established based on the actual intersection, and the three control methods proposed in this paper are simulated and analyzed. The comparison of delay shows that the control effect of fuzzy neural network control is the best, the delay relative timing control is 20% -30%, fuzzy control is the second, and relative timing control is 10% -15%. Finally, the variation law of vehicle queue length and timing scheme in the process of fuzzy neural network control simulation is analyzed in detail, the relationship between queue length and timing result is discussed, and the scientific nature of the control method is verified.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.54

【参考文献】