Capacity is one of the most critical concepts in traffic engineering practice. Microsimulation models generally ignore this concept primarily because they are based on car-following behaviour of individual vehicles. This paper attempts to establish a clear connection between the operation of microsimulation models and traffic engineering theory at the intersection level. PARAMICS queue discharge flow rates at a hypothetical signalised intersection are first compared with exponential models proposed by Akcelik. Fitting of microsimulated virtual data sets to queue discharge flow and headway exponential models is done using non-linear regression. This process enables identification of key microsimulated queue discharge flow parameters such as maximum saturation speed, maximum saturation flow, minimum saturation headway and vehicle spacing. Using discharge flow parameters determined from non-linear regression, vehicle acceleration, speed and distance profiles are derived and compared to polynomial acceleration models adopted in SIDRA software. Car following model parameters are then adjusted in order to achieve more realistic saturation flow rates similar to those observed in the field. The paper explains how changing a set of car-following model parameters affects properties of queue discharge flows. It also confirms that observed parameter values (e.g. driver reaction time) may not be appropriate when applied in the context of car-following model algorithms. (a) For the covering entry of this conference, please see ITRD abstract no. E217099.
Samenvatting