Trip Travel Time Distribution Prediction for Urban Signalized Arterials

Travel time prediction is a challenge, especially if we consider urban trips. For freeways, well-known models for traffic flow and speeds are applicable, e.g., based on physical models inspired by hydrodynamic or statistical models ranging from more conventional to more advanced AI (artificial intelligence) approaches; while for urban trips the traffic flow models are more complicated because, next to vehicle-vehicle interaction, the influence of traffic signals must also be modeled. In this paper, a trip travel time distribution model for urban roads with fixed-time controlled intersections is proposed. Based on the proposed model, a trip travel time distribution prediction procedure is discussed, which considers time-varying demand and traffic control schemes. The model-predicted results are further compared with those from VISSIM simulation data. It shows that the proposed trip travel time distribution prediction model can perform well both for undersaturated conditions and oversaturated conditions.