An Empirical Study on Parameters Affecting Traffic Stream Variables Under Rainy Conditions

Inclement weather condition such as rainfall greatly affects the traffic stream variables. To study the impact of rainfall on vehicle mobility, fine-grained data is required which is provided by the Intelligent Transportation Systems (ITS) infrastructure and weather stations. As the installation and maintenance of ITS infrastructure are costly, the majority of traffic data collection is carried out using cost-effective alternate sources such as cellular networks and GPS probes. These alternate sources of traffic data provide sparse, incomplete, and erroneous information. To overcome the issue of data sparsity, we propose a mechanism to generate fine-grained synthetic traffic data using the Simulation of Urban Mobility (SUMO). It is found that a large number of weather parameters affect traffic mobility in the region. We design a generic empirical model that captures the impact of rainfall intensity, road type, friction, visibility index, and the time of day on the traffic stream variables. The designed empirical model is integrated into the Krauss car-following model of SUMO. The simulation model is validated using the traffic data reported in the literature under rainy conditions. We find that the synthetic data generated using the proposed empirical model matches well with the traffic data reported in the literature.