The data presented here show that human drivers apply a discrete noisy control mechanism to drive their vehicle. A car-following model built on these observations, together with some physical limitations (crash-freeness, acceleration), led to non-Gaussian probability distributions in the speed difference and distance which are in good agreement with empirical data. All model parameters have a clear physical meaning and can be measured. Despite its apparent complexity, this model is simple to understand and might serve as a starting point to develop even quantitatively correct models.
@article{arxiv.physics/0601058,
title = {How human drivers control their vehicle},
author = {Peter Wagner},
journal= {arXiv preprint arXiv:physics/0601058},
year = {2009}
}