Drivers' behaviour modelling for virtual worlds a multi-agent approach

Abstract

In this paper we present a study that looks at modelling drivers' behaviour with a view to contribute to the problem of road rage. The approach we adopt is based on agent technology, particularly multi-agent systems. Each driver is represented by a software agent. A virtual environment is used to simulate drivers' behaviour, thus enabling us to observe the conditions leading to road rage. The simulated model is then used to suggest possible ways of alleviating this societal problem. Our agents are equipped with an emotional module which will make their behaviours more human-like. For this, we propose a computational emotion model based on the OCC model and probabilistic cognitive maps. The key influencing factors that are included in the model are personality, emotions and some social/personal attributes. © Springer International Publishing Switzerland 2013.