The aim of this article was to explore the potential benefits of behavioural optimum analysis to the understanding and prevention of traffic risks and accidents. Based on the general assumption that drivers try to optimize their behaviour not for traffic safety but for their own (subjective) preferences, we developed a formal model of cornering behaviour. Within the framework of a physical description of the scenario we integrated the assumed driver preferences in terms of an evaluative function which maps the continuum of possible cornering speeds into a single dimension - the so called reinforcement value. Using a numerical algorithm we optimized this hypothetical function for maximum reinforcement value, resulting in a behaviourally optimal velocity. In a second step we compared this value to what can be considered as a maximum safe velocity and found that in our model the distance between preferred speed and safe speed depends on the radius of the curve under consideration. This implies that people do not have a general tendency to drive at safe speed or to drive too fast, but that the safety of their speed varies systematically with the physical environment. The limitations of the presented work lie, of course, in its purely hypothetical character. Both the evaluative function and the parameters used in the above analysis do not claim to model real driving behaviour but are to be understood as mere examples to demonstrate the potential usefulness of our general approach. The aim was to show how a behavioural maximization model can be used to identify unrecognized safety factors in a traffic scenario. Since up to now there has been no empirical validation of the model, the conclusions should be interpreted with care. Moreover, the physical representation of the traffic scenario is rather simplified - factors like other cars, or the properties of the vehicle used by the driver may in fact play an essential role regarding the behavioural optimum. Apart from that, our optimization model is static, that is, it does not model the dynamics of driving. Neither does it take into account that preferences may change depending on the context. To incorporate these features the use of differential equations and advanced formalisms with a greater scope would be necessary (Borgstede, Schicke, Eggert & Goltz, 2011). However, the current model suffices to demonstrate the basic principles of our approach - without making things unnecessarily complicated. Despite its hypothetical character the above modeling approach offers a useful contribution to the analysis of driving behaviour in general and, more specifically, traffic safety. Focusing on behavioural optima which are the result of subjective preferences, our approach opens a new perspective of applying psychological principles to traffic behaviour: Formal models of driving behaviour need not restrict themselves to understand the cognitive processes which make driving possible, but can investigate motivational factors in a scientifically rigorous way. The resulting predictions (e.g. concerning optimal curve radius to minimize drivers' tendency to drive too fast) can be applied to make people "voluntarily" drive in a safe way - without them even noticing it. Thus, the theoretical analysis in terms of behavioural optimization can lead to new insights which may give new impulses to enhance traffic safety.
Analyzing Traffic Safety Factors using Behavioural Optimization Models
Fahrermodellierung in Wissenschaft und Wirtschaft, Berliner Fachtagung Fahrermodellierung, 4 ; 57-64
2013
8 Seiten, Bilder, Quellen
Conference paper
English
Traffic safety evaluation using behavioural criteria
Automotive engineering | 1993
|93SF094 Traffic Safety Evaluation Using Behavioural Criteria
British Library Conference Proceedings | 1993
|British Library Conference Proceedings | 2002
|