An Improved Moving Observer Method for Traffic Flow Estimation at Signalized Intersections

With the deployment of partially and highly automated vehicles, the automotive industry is greatly increasing its influence on road traffic. In order to ensure a positive influence of automated vehicles on traffic efficiency as well as traffic safety, simulations are broadly used for the development and testing of the required functions. Since these simulations are applied to evaluate the behavior of an automated driving function in the real world, an exact representation of the real world in the simulation is essential for the validity of the generated results. Therefore, there is a need for methods with which certain parameters of real-world situations may be quantified and applied to a simulation. In this work, we propose an approach to measure traffic flow and estimate the traffic state in a network based on extended floating car data. For this purpose, the data concerning the movement of the tracked vehicles is combined with the data regarding surrounding traffic gathered by the vehicles' sensors. The aim of this combination is to achieve an accurate traffic observation on urban as well as rural roads with a minimal number of test vehicles gathering the data. The application of the method to simulated traffic results in an accurate estimation of the traffic volume. The functionality is also demonstrated based on a limited sample of real-world test data.