This report presents the conceptual framework for the development of a traffic flow model that is capable of modelling the potential influence on traffic flow dynamics of Advanced Driver Assistance (ADA) systems during the emergence, presence and dissolution of congestion on motorways. We aim at improving the modelling of empirically observed phenomena in congested traffic by explicitly incorporating more refined knowledge of driver behaviour into the traffic flow model. For the individual driver model a mode-based system dynamic model was applied to the task of driving. The model explicitly accounts for typical human characteristics such as delayed response, erroneous perception etc. and uses behavioural knowledge from theory and experiments. For the traffic flow model the so-called mesoscopic or (gas-) kinetic modelling approach was selected. The 'driver psychology' is accounted for by defining the level of attention as a function of the traffic conditions. The performance of the individual driver - and therefore the parameters used in the micro-macro link - are attention-level-dependent. This means that the drivers' mental state is explicitly modelled on both micro and macro levels. (A)
Abstract