Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (13): 214-216.DOI: 10.3778/j.issn.1002-8331.2009.13.063

• 工程与应用 • Previous Articles     Next Articles

Traffic condition recognition based on probability neural network

GUO Geng-qi1,2,CAO Cheng-tao1,2,XU Jian-min1   

  1. 1.South China University of Technology,Guangzhou 510640,China
    2.Guangdong Communication Polytechnic,Guangzhou 510650,China
  • Received:2008-12-13 Revised:2009-02-13 Online:2009-05-01 Published:2009-05-01
  • Contact: GUO Geng-qi

城市道路状况概率神经网络判别方法

郭庚麒1,2,曹成涛1,2,徐建闽1   

  1. 1.华南理工大学,广州 510640
    2.广东交通职业技术学院,广州 510650
  • 通讯作者: 郭庚麒

Abstract: A traffic condition recognition method based on floating car data is proposed by analyzing Probability Neural Network(PNN) and Global K-means algorithm.The related factors of traffic condition and the collection method of floating car data are presented.Considering the influence of traffic control intersection delay to travel time,a probability neural network classifier is designed using Global K-means algorithm and applied to the recognition of traffic condition with floating car data.The experiment results show that the method can recognize traffic condition well,which can reflect traffic condition better than that without considering traffic control intersection delay.

Key words: traffic condition, floating car, probability neural network, Intelligent Transport System(ITS) information platform

摘要: 针对移动交通流检测信息的特点,在分析概率神经网络与Global K-means聚类算法的基础上,提出了一种基于移动交通流检测信息的城市路况概率神经网络判别方法。通过分析路况的相关因素,同时考虑信号控制交叉口红灯对车辆行程时间延误的影响,利用Global K-means算法改进的概率神经网络对探测车采集的实时交通信息进行处理,进而得出城市的道路状况。应用结果表明该方法能够有效地判别和跟踪道路状况的变化,比不考虑交叉口红灯的影响时能够更准确地反映城市道路的路况信息。

关键词: 城市路况, 探测车, 概率神经网络, 智能交通系统信息平台