A Cognitive Tool to Predict Individual Accident Proneness of Vehicle Drivers

Ying Wang,Licheng Xue,Jing Li,Kan Zhang,Xianghong Sun
DOI: https://doi.org/10.1061/41039(345)205
2009-01-01
Abstract:Cognitive abilities are believed to be highly correlated with driving safety (Anderson et al., 2005). Developing cognitive tasks to predict individual traffic accident proneness could benefit both driver screening and traffic accident prevention. In this study we designed four tasks to measure vehicle drivers' cognitive abilities consisting of attention span, attention distribution, spatial working memory span, multiple object tracking and speed estimation. Significant correlation was found between these cognitive indexes and the drivers' crash frequency in a simulated driving task. Then the drivers were divided into high accident-prone group and low accident-prone group based on their crash frequency. Discriminant analysis using cognitive indexes as discriminant variables revealed an up to 94.4% highly accurate classification pattern, which approved the validity of our cognitive tool as a useful predictor of individual driver's accident proneness.
What problem does this paper attempt to address?