单交叉路口交通信号模糊控制系统的研究车辆工程专业论文.docxVIP

中南大学硕士学位论文 中南大学硕士学位论文 摘要 摘要 本文针对当前交叉路口交通控制信号灯配时固定不变的问题，提 出了一种基于短时交通流量预测的模糊控制方法，用于适时控制交通 信号的相位长度及相位时序，以满足交通流量高峰期和平时流量的通 行需求。并在此方法的基础上，完成了单交叉路KI交通信号模糊控制 器的研制。本论文完成了如下工作： 1．对城市道路交通流量的短时预测方法进行了研究，基于遗传算 法对神经网络预测模型进行改进，建立了GA—|BP网络预测模型，并 且通过仿真实验论证了改进后的模型在预测精度上有很大的提高，同 时预测过程所需时间减少，可用于在线控制。 2．设计了四相位单交叉路口交通信号的模糊控制模型，提出了一 种以当前相位、后继相位的车辆等待数和未来短时预测流量来决定信 号配时的多相位模糊控制方法，同时提出单交叉口的交通流量模型， 使其能同时反映轻度交通和重度交通，能够描述排队长度。最后通过 仿真验证了该控制方法在实际应用中的可行性和优越性。 3．给出了以AT89S51单片机为主机的控制系统的硬件和软件的设 计。在以上基础上以长沙市五一路口实际流量作为实验数据，进行了 模拟实验。实验结果表明该控制系统能达到预期效果。 关键词：城市交通控制，单交叉口，神经网络预测，模糊控制，单片 机 中南人学硕上学位论文 中南人学硕上学位论文 ABSTRACT Abstract In view of all the problems in traffic，a kind of distributed fuzzy logic controller has been presented in this dissertation，with which a kind of short—term traffic prediction method based on BP neural．Study on the simulation of the fuzzy control method and the traditional time method (simulated by Matlab)．And the results show that this fuzzy logic control method overmatches traditional time method．It’S really one of the efficient methods to solve the problem of urban traffic． 1．Study on short—term traffic prediction has been done．Improve the neural network forecast model based on the genetic algorithm and establish the GA-BP network forecast model．The simulation prove that GA—BP network forecast model obtain a big enhancement in the forecast precision． 2．With the traffic flow model established，a kind of traffic controller based on GA—BP and fuzzy logic，which is combined with GA—BP short-term traffic flow predi ction and the phase sequence optimization module，has been designed．According to the needs of this dissertations，analyzes the correlations between the macroscopic traffic flow theories and microscopic car-following models．Form a new traffic flow model suitable for the study of this dissertation Select the special streets in Changsha city as the studying object． 3．The hardware realization and the software program with AT89S5 1 chip microcomputer are given．According to the