A Psycho-Social Agent-Based Model of Driver Behavior Dynamics

Abstract

This paper suggests a psycho-social agent-based model, referred to as Holistic-Emergent Social Interaction-Oriented Dynamics (HESIOD) model, to simulate the drivers' behavior dynamics under various types of interaction among vehicles. The HESIOD model allows representing the heterogeneity and dynamical processes involved in such control dimensions as risk assessment and time responsiveness of driving behavior (controlled dimension). It is shown that highly differentiated states may arise, such as fixed point, periodicity and transient chaos. The dynamical state is found to be mostly affected by the degree to which the control dimensions of neighboring vehicle drivers depart from each other, and the topology of interaction among drivers. In contrast with the aggregate statistical-probabilistic models, this agent-based model can offer valuable insights into the role of both cognitive processes and interactions of drivers on their actual driving behavior. The findings may have useful implications for improving the level of service, safety and security in roads.