基于动态模糊神经网络的交通信号智能控制研究

发布时间：2018-11-02 20:13

【摘要】：进入21世纪以来,随着我国经济和科技力量的突飞猛进,道路上的汽车数量爆发式增长,城市交通拥堵越来越严重,进一步引发了环境污染、能源浪费、交通事故等一系列问题,尤其雾霾天气给人们带来很大的困扰。城市交通拥堵主要发生在道路交叉口处,由于传统的交叉口交通信号控制方式存在对绿灯时间分配不合理的问题,有时会造成不必要的拥堵。一个合理的控制系统,对改善交通现状具有深远的影响。在道路交通网中,交通流是非线性的、实时的和变化无常的,针对交通流的这种特性,可以把智能控制方法应用于城市交通信号控制中。本文主要研究了城市中比较特殊的五岔路口和相邻交叉口,把模糊系统和神经网络结合起来应用到交通信号控制中,以降低车辆平均延误为目标,实现五岔路口绿灯时间的合理分配。首先,简单介绍了交通信号控制基本参数、交通流的统计分布、道路交叉口交通质量评价指标体系和交通流检测。其次,将模糊控制应用到五岔路口的交通信号控制中,实现五岔路口的智能控制。在车流低峰期和高峰期两种情况下,分别用定时控制和模糊控制仿真研究五岔路口控制,仿真结果显示了模糊控制的优势。再次,针对交通信号定相序控制浪费绿灯时间和频繁切换相位的问题,引进动态模糊神经网络理论,实现五岔路口多相位变相序动态控制。以赤峰市松州路、振兴大街和临潢大街所形成的五岔路口为例,对其进行了仿真研究,验证了多相位变相序动态控制方法的性能。最后,研究分析了相邻交叉口的协调控制。对于中间路段距离小的相邻交叉口,其关联性比较强,在控制中考虑了中间路段驶入和驶离交叉口的车流对绿灯相位的影响,仿真说明协调控制比普通的孤立交叉口控制更加合理。
[Abstract]:Since the beginning of the 21st century, with the rapid development of our economy and science and technology, the number of cars on the roads has exploded, and the traffic congestion in cities has become more and more serious, which has caused further environmental pollution and energy waste. Traffic accidents and a series of problems, especially the haze weather to people a great deal of trouble. Urban traffic congestion mainly occurs at road intersections. Because of the unreasonable allocation of green time in the traditional traffic signal control mode at intersections, it sometimes causes unnecessary congestion. A reasonable control system has a profound impact on improving the traffic situation. In road traffic network, traffic flow is nonlinear, real-time and fickle. In view of this characteristic of traffic flow, intelligent control method can be applied to urban traffic signal control. This paper mainly studies the special five fork junctions and adjacent intersections in the city. The fuzzy system and neural network are combined into the traffic signal control, with the aim of reducing the average vehicle delay. Realize the rational allocation of green light time at five fork junctions. Firstly, the basic parameters of traffic signal control, the statistical distribution of traffic flow, the evaluation index system of traffic quality at road intersection and the detection of traffic flow are briefly introduced. Secondly, the fuzzy control is applied to the traffic signal control of the five bifurcations to realize the intelligent control of the five bifurcations. In the case of low peak period and peak period of traffic flow, the fuzzy control and timing control are used to simulate the five-fork intersection control, respectively. The simulation results show the advantages of fuzzy control. Thirdly, aiming at the problem of wasting green time and switching phase frequently in phase sequence control of traffic signal, the dynamic fuzzy neural network theory is introduced to realize multi-phase variable phase sequence dynamic control at five junctions. Taking the five fork junctions formed by Songzhou Road, Zhenxing Street and Linghuang Street in Chifeng as an example, the simulation study is carried out to verify the performance of the multi-phase variable phase sequence dynamic control method. Finally, the coordinated control of adjacent intersections is studied and analyzed. For the adjacent intersections with small distance from the middle section, the correlation is relatively strong. In the control, the influence of the traffic flow of the middle section entering and leaving the intersection on the green phase is considered. The simulation results show that the coordinated control is more reasonable than the common isolated intersection control.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.5

【参考文献】