A method to predict travel time in large-scale urban areas using Vehicular Networks

The suggestion of route with shortest travel time avoids unnecessary traveling by cars. Achieving this goal not only reduces air pollution, but also it saves time and money. In recent years, several methods have been proposed to predict routes travel time. Vehicular networks are considered as an efficient approach to traffic management applications. Since it takes time to cover all the streets and equip all vehicles with the network connection equipment, in the present study, a method has presented and simulated in which vehicles in the parts of the roads that are not covered by vehicular networks, be able to predict travel time of their routes according to past traffic information. So two paradigms of Artificial Intelligence (AI) have been used for this purpose. This study overcomes two important constraints in the previous studies. First, in most of the previous studies, in order to predict travel time using vehicular networks, the installation RSUs was necessary for all intersections. In the proposed method, this restriction is removed. Secondly, the instantaneous and accurate data obtained from the vehicular networks were used for data training of the artificial neural network. In previous works, training data was collected with old equipment such as detectors, traffic monitoring cameras, sensors, etc. This equipment can inform about the status of traffic parameters in the installed area and do not cover the roads network entirely. Finally, the accuracy of prediction of travel time with two models of AI is evaluated and a model with less prediction error is used for simulation. The results show that the proposed method can predict travel time of routes with the desired precision and it can conduct cars to the route with the shortest travel time.