Applications of Traffic Flow Predictions and Simulations in Urban Mobility with Connected Automated Vehicles and Advanced Traffic Management Systems

Urban mobility faces serious problems in daily life caused by increasing population density and traffic demand, leading to congestion at urban roads with limited capacity. Adding to this, the future era of connected and automated vehicles in urban mobility will influence the traditional urban traffic stream and traffic infrastructure. Real-time traffic solutions using the strategies in Intelligent Transportation Systems (ITS) like Advanced Traffic Management Systems (ATMS) and Advanced Traveller Information Systems (ATIS) are being analysed and implemented extensively for reduction or avoidance of congestion. These real time strategies can be used along with traffic flow prediction models which subsequently serve to inform both the network and the traveller about current and future traffic scenarios. Hence travel delays and vehicle idling can be reduced, leading to reduction in the queue length, waiting time and fuel consumption. This research work includes prediction of real time traffic flow at a busy intersection in Duisburg using a state of art time series prediction model : Fb-Prophet. The predicted traffic flow will be used to build a real time traffic simulation scenario using SUMO. A python-based tool, Flow generator, has been developed to create vehicle flow in SUMO from the prediction results. The suggested traffic flow prediction model can also include the effect of external factors influencing traffic flow such as the day of week, peak/off-peak hour, holidays and severe weather (rainfall, snow). Hence, the predicted traffic flow for every lane will be given as input for simulation in SUMO which can be further used to analyse the impact of connected vehicles and current traffic control system.