Estimating the urban traffic state with limited traffic data using the MFD

Abstract: Urbanization leads to an increase of traffic in cities. The Macroscopic Fundamental Diagram (MFD) suggests to describe urban traffic at a zonal level, in order to measure and control traffic. However, for a proper estimation, all data needs to be available. The main question discussed in this paper is: How to derive a network-wide traffic state estimate? We follow up on literature suggesting to base the operational estimate on the speed of limited sample cars sharing floating car data (FCD). We propose an initial step by constructing an MFD based on FCD, which is then used in step 2, the operational traffic state estimation. For operational traffic state estimation, i.e., the real-time traffic state estimation, the penetration rate is unknown. For both steps, we assess the impact of errors in the estimation. In light of the errors, we also formulate an indicator which shows when the method would yield insensible results, for instance in case of an incident. The method has been tested using microsimulation. A 26% error in the estimated average density is found for a FCD penetration rate of 1%; increasing the penetration rate to 30\% reduces the error in estimated average density to 7%.