A hierarchical risk model for traffic participants

Abstract

As traffic participation is inherently a risky activity, traffic psychology has generated a great number of so-called risk models, i.e. models in which the risk concept plays a major role. Three of these models are attracting a great deal of attention these days: Näätänen and Summala's ‘Model of drivers’ decision making and behaviour', Wilde's ‘Theory of risk homeostasis’ and Fuller's ‘Threat-avoidance model of driver behaviour’. All three models emphasize motivational aspects with regard to risk and they claim to be generally applicable to a large array of traffic situations. In an attempt to use these models for quantitative predictions in a concrete example (an overtaking manoeuvre), we found that many model components had not been defined at all, or had been defined only partially, or in a contradictory fashion. We have therefore developed our own model which allows quantitative calculations in terms of behaviour alternatives, subjective probabilities of events, and utilities of the outcomes of behaviour alternatives. The concept of risk is more sharply defined as well. Further, the model explicitly takes into account that traffic tasks may be conceived as hierarchically ordered in strategic, tactical and operational task levels.