Paper Details

TRB 91st Annual Meeting (January 22-26, 2012)

Event Number:	493
Event Title:	Research in Traffic Flow Theory and Characteristics, Part 4 (Part 1, Session 403; Part 2, Session 433; Part 4, Session 493)
Event Date:	Jan 24 2012 8:30AM- 10:15AM
Event Location:	Marriott, Salon 2
Event Description:
Event Agenda:

Toward a Stochastic Psychospacing Model (12-2343)

Advanced vehicle control and safety systems have great potential in improving safety, traffic flow and environmental issues. Human factors can be assumed to play a substantial role in the efficacy of these systems. Human factors however lead to substantial differences in driving behavior within as well as between drivers. In order to determine the efficacy of these systems it is therefore crucial that mathematical models of driving behavior capture this stochasticity caused by human factors. A complicating factor is however that in the past driving behavior has predominantly been modeled through deterministic models. In order to capture this stochasticity in driving behavior in this contribution a first step is taken towards modeling of car-following through a Bayesian network approach based on psycho-spacing theory. A Bayesian net was set up and parameters were learned to the net using empirical trajectory data. From the marginal and joint probabilities it follows that the model performs relatively well in predicting car-following behavior. The contribution is concluded with a discussion as well as recommendations for future research.

Authors
     Hoogendoorn, Raymond Gerard, Delft University of Technology, Netherlands
     Hoogendoorn, Serge , Delft University of Technology, Netherlands
     Brookhuis, Karel A., Delft University of Technology, Netherlands