Traffic Forecasting at Non-Detector Roads Based on City Road Network

Abstract

This paper aims to predict the traffic parameters of non-detector roads in a network. The traffic parameter data are collected by a small number of detectors installed on the some of the roads. In a clustering analysis we consider three traffic parameters: traffic volume, average speed and time occupancy ratio. This paper uses the dissimilarity dynamic matrix to process the array data consisting of the above mentioned three dimensions, to obtain the value of traffic parameters of all roads in a regional network. Furthermore, we have chosen a small network in Xi’an to check if this method works. These traffic parameters can be used to predict traffic status to provide the basis for traffic guidance system and the traffic management system.