Year of publication:
Licentiate thesis, Department of Applied Mechanics, Chalmers University of Technology
Inattention is one of the most common factors contributing to road crashes. However, the basic mechanisms behind inattention and how it leads to crashes are still relatively poorly understood. The objective of the present thesis was to develop a testable model of attention selection in driving which could serve as the basis for a scientifically grounded taxonomy of drivers’ attention failures. A key starting point for the model development was that inattention needs to be understood within a broader framework of adaptive driver behaviour. A proposal for such a framework was outlined as the basis for the subsequent model development. At the core of the proposed model is an attention selection mechanism based on the biased competition hypothesis, which states that attention selection occurs through local competitive interactions which are biased top-down and/or bottom-up. The model is rooted within the embodied cognitive science tradition and adopts an activation dynamics-, rather than an information processing, metaphor. Still, the main concepts are largely compatible with traditional information processing models of attention, such as Multiple Resource Theory. The model was validated against existing empirical work on driver inattention, in particular the two studies reported in the appended papers, which investigated the effects of visual and cognitive secondary tasks on driving performance, behaviour and state. It was demonstrated how the model offers novel explanations for several observed phenomena not which do not seem to be accounted for by existing models. Moreover, a number of novel predictions were generated which could be tested in future empirical studies. It was also demonstrated how the model can be used as the basis for a taxonomy of attention-related failures and factors in driving.
More information / Full text:
Traffic Safety Analysis
FICA 2 - Factors Influencing the Causation of Accidents and Incidents