For the analysis of the existing and planned segments of the traffic system, a traffic modelling is used, and the choice of models depends on temporal and spatial limits of a model and the context of application of modelling results. Functioning of a traffic system is under the influence of variable human behaviour. Researches show that the behaviour of a driver is, among other things, territorially and culturally conditioned. Accordingly, there is no universally applicable model, so the adjustment of modelling to local characteristics of traffic
system and its users is a necessary prerequisite for the use of any traffic model. The success of modelling of urban traffic networks and its segments is interrelated with successful modelling of critical network segments. According to a number of criteria, the most critical point of every traffic system is the intersection. The simulation models are very useful tool for the analysis of existing critical network segments and prediction of traffic conditions on existing and planned intersections. However, it is questionable whether they can be expected to give realistic modelling results that can be applied in the methodology, analysis and design of intersections in local conditions. Within this paper, a special attention has been paid to the microsimulation modelling of urban roundabouts. Accuracy of modelling results has been established by the comparison of travelling times between measuring points gathered in calibrated and uncalibrated VISSIM microsimulation models and the ones gathered in the field.

travel time, microsimulations, vissim, roundabouts