For the design of a vehicle control algorithm to monitor and correct longitudinal driving behavior, it is essential to have a good insight into thedifferent parameters that determine this behavior. In this research, eleven systems and control related parameters are identified, which are related to time headway (THW), time-to-collision inverse (TTCi), and the switch time between accelerator release and brake activation. These parameters are used to determine and distinguish between dissimilar types of longitudinal driving behavior, in terms of driving and driver characteristics. The efficient K-means clustering algorithm is used for the classification of longitudinal driving behavior. Driver behavior experiments were carried out that involved forty-five participants. The results of the study show that four main determinants of longitudinal driving behavior can be distinguished using measurable parameters, with the indicated opposite extreme values: prudence (aggressive vs. prudent); stability (unstable vs. stable); conflict proneness (risk prone vs. risk infrequent); and skillfulness (non-skillful vs. skillful).
Abstract