Understanding drivers' behaviour is essential to design traffic infrastructures with high degree of safety, capacity and comfort. The lack of consideration of drivers' behaviour, referring to a bad model, or even worse the blind use of traffic signals that drivers will never respect, may lead to conditions of extreme unsafety such as under/over rating the capacity of the infrastructure. The study of users behaviour in a few tapered on-ramps in Italy pointed out how behaviour may be mispredicted, and what consequences wrong evaluations may lead to: rough schematisations that do not consider the interaction between the flows, or that regard interaction as a one way phenomenon (drivers on the merging lane are influenced by drivers on the main lane), are not acceptable. When considering the merging manoeuvre on the whole, four different models have been created or adapted from the classical one: (1) car following model; (2) lane changing; (3) behaviour on section just before the merging lane; and (4) behaviour on tapered section. These four models have been integrated in a micro simulation program written in Visual Basic. Such program seems to describe very well the observed behaviour, and may be used to predict infrastructures functioning and get design prescription. Simulations give indication in term of capacity, delay and comfort taking into account as variables the typology of the merging lane (tapered ¡ parallel), geometry of the main street / on ramp (number of lanes, visibility, length of the merging lane etc) flow rates, speeds, users typology distribution (aggressive, prudent etc). Results show great differences and give interesting suggestions to be used at design level. (A) For the covering abstract of the conference see ITRD no 207828. The reprints are also available at the web - http://www.vti.se/pdf/reports/K18APart1.pdf; http://www.vti.se/pdf/reports/K18APart2.pdf and http://www.vti.se/pdf/reports/K18APart3.pdf.
Samenvatting