A Control-oriented Macroscopic Traffic Flow Model for Urban Diverse Intersections

Abstract

Traffic flow on roads is a non-linear, stochastic phenomenon, with complex interactions between vehicles. This paper discussed the dynamic nature of urban intersections, and presents a novel traffic flow evolution model at a time scale and of a level of detail suitable for on-line estimation, simulation and control. The intersection is considered as interconnected components of urban road network. The macroscopic model proposed here extends the platoon-dispersion model by redefining inflow and outflow vectors, and by also specifying the control vector. Two variables were added to the model, i.e. the turning proportion and the lane assignment scheme. The turning proportion is an obvious variant, and yet the lane assignment is time-varying in some especial condition. Simple stochastic matrix equations described the macroscopic traffic behavior of each movement. This will allow the simulation of quite large road networks by composing many intersection models. The model is validated over real traffic data with abrupt changes in the demand of flow.