Data fusion algorithms for density reconstruction in road transportation networks

This paper addresses the problem of density reconstruction in traffic networks with heterogeneous information sources. The network is partitioned in cells in which vehicles flow from their origin to their destination. The state of the network is represented by the densities of vehicles in each cell. Density estimation is of crucial importance in future Intelligent Transportation Systems for monitoring, control, and navigation purposes. However, deploying fixed sensors for this purpose can be very expensive. Therefore, most of fixed sensors networks are rather sparse. On the contrary, recent technologies have enormously increased the availability of relatively inexpensive Floating Car Data. A data fusion algorithm is then proposed to incorporate the two sources of information into a single observer of density of vehicles. The efficiency of the proposed algorithm is shown in a real scenario using data from the Grenoble Traffic Lab fixed sensor network and INRIX Floating Car Data on the Rocade Sud in Grenoble.