Prediction of a Street Travel Time in GIS by Using Periodic Parameters in Traffic Statistical Data

Nowadays, the problem of transportation and traffic has become a social challenge in many countries. With the advancement of science, a favorable environment for an intelligent and purposeful management is achieved in order to improve productivity and increase efficiency of network traffic. In the field of information technology, Geographic Information Systems are very important in optimizing the performance of transportation systems. In this context, the capabilities of network analysis in GISs such as shortest path computing could be very useful.
Different criterions have ever considered for shortest path analysis in GIS. Distance, travel time, path comfort, path beauty and etc., are some of these criterions. Travel time criteria has some random continues variations, because of its relation to traffic. Therefore, the travel time of each street must be calculated based on its geometric and traffic momentarily.
In this paper we have developed a model for this purpose that is based on wavelet transform and least square method. This model uses the statistical information of travel time in previous days for prediction of travel time in future. Finally for testing the developed model a case study have been explained in this paper.