Evolutionary approach for the traffic volume estimation of road sections

In recent years, a vast amount of real time traffic information is collected and provided to the travelers as a part of Intelligent Transportation Systems. These information is collected using sensors or detectors etc. set on the road sections and is utilized by car navigation devices to guide the travelers efficiently in the road network, or used to predict the future traffics. However, sometimes these information is not available for all the road sections in the road network. Generally speaking, road sections are classified into several different categories and currently real time traffic information is available only in road sections in major categories. In this paper, a genetic algorithm approach is proposed to estimate the traffic volume in road sections without the traffic information, where estimation is done using the known traffic volume information of the road sections. The proposed method is evaluated under static environments using a grid road network with various unknown rates of traffic volumes. Experimental results show that the proposed method can estimate the unknown traffic volume using only the known traffic volumes.