A study on the behaviour of microscopic car-following models in urban settings

Abstract

This thesis carries out empirical analysis and compares four common continuous microscopic models in typical urban scenarios against each other and real human driver behaviour. The best performing model is found to be unstable travelling at the speed limit in a platoon, thus a variant of this model is proposed and implemented. A heterogeneous network uses this continuous model and a discrete microscopic model to identify a distinct density signature where the models switch over.