This paper describes the results of a study into how realistically traffic models can describe traffic movements and what the effects of shortcomings in the models are on emission calculations, e.g. in the evaluation of ITS measures. Data were collected at a signalised intersection in Rotterdam, using camera images to construct vehicle trajectories. The observed traffic patterns were compared to traffic patterns simulated with the micro-simulation models VISSIM and AIMSUN. Clear differences were found in free driving behavior, as well as in the deceleration and acceleration patterns for vehicles that have to stop for a red light. The differences lead to both under- and overestimations of emissions. These errors can have substantial consequences when evaluating measures that influence the driving behavior at intersections. Several suggestions for improvements are given.
Abstract