Modeling Human Aspects of Driving Behavior
摘要
The driving characteristics described by the microscopic models of the previous chapters correspond, from a formal perspective, to semi-automated driving as realized by adaptive cruise control (ACC). On the one hand, human drivers are less efficient than ACC systems since reaction times, attention time spans, and estimation errors play a significant role. On the other hand, humans can take into account more input stimuli than acceleration controllers, for example: brake lights, turning signals, next-nearest neighbors, and external conditions. Moreover, in contrast to the present-day ACC systems reflected by the previous models, they can anticipate the situation for the next few seconds. All these specific human aspects will be formulated in terms of psycho-physiological extensions to the previous car-following models, in particular Gipps’ model and the Intelligent Driver Model. Another class of psycho-physiological models explicitly take into account finite perception thresholds leading to sudden changes in accelerations whenever the difference from the ideal acceleration becomes significant. We present action points and the Wiedemann-74 and Wiedemann-99 models as representatives of this model class. Particularly, we present full and consistent longitudinal specifications of the two Wiedemann models which are difficult to access elsewhere.