Development of an on-ramp model for use in microscopic traffic simulation models

Abstract: Traffic simulation models have become more and more important for the traffic industry over the past few years. This thesis enables the microscopic traffic simulation model of Johan Olstam to include merging areas in its simulation runs and thereby takes the model one step further on its way to realistically simulate vehicles in all kinds of traffic situations. A state-of-the-art literature review as well as an analysis of Olstam’s model are used for identifying possible weaknesses of the current merging behaviour model. Based on this knowledge, an improved merging behaviour model is developed. This model has to perform merging manoeuvres as realistically as possible since the microscopic model’s purpose is to simulate surrounding vehicles in driving simulators. By implementing new concepts in the gap selection, gap acceptance, calculation of the merge acceleration and courtesy deceleration as well as adding new functions allowing cooperative lane changes and merge anticipation, the original merging behaviour model has been improved significantly. The enhanced model is analysed and verified by comparing its results to real world data and data of the original model, respectively, as well as by studying simulated vehicle trajectories.