Modelling traffic dynamics in motorway networks

Due to the rapid growth of traffic density, the necessity to increase the operational efficiency and capabilities of intelligent transportation systems (ITSs) has led to the development of various traffic modelling theories. The Lighthill-Whithman and Richards (LWR) model [?, ?], uses fluid based partial differential equations to capture traffic dynamics along continuous stretches of road. In contrast to the LWR model, the artificial neural network model [?, ?] utilizes historical observations of traffic flow-rates to forecast flow-rate locally. This paper aims to introduce a new hybrid macroscopic model which combines the complementary features of the LWR and artificial neural network models, to effectively simulate traffic flow in road networks. The model developed in this paper demonstrates the ability to, within a certain degree of accuracy, forecast traffic flow in a road network that includes junctions and continuous stretches of road. Furthermore, the proposed model offers an appropriate trade-off between accuracy and computational complexity, therefore it is suitable for real time applications.