Evaluation of Classification Models for Goals in Visual Search Using Driver Segments Based on Cognitive Characteristics

Drivers continuously shift their visual attention between the road and the dashboard to search information which they need for driving. According to Cooper (2013), their driving performance is affected by cognitive aspects; therefore, it is necessary to consider drivers’ cognitive characteristics and abilities in a driving context. Existing studies on the evaluation of drivers’ performance are focused on gender or age difference (Oltedal, 2006; Lonczak, 2007). In the precedent study (Jo & Lim, 2013), 114 users were clustered into eight segments, and for three selected segments, nine users were recruited for eye movement data gathering during a dual task of driving and visual search. In this study, logistic regression classification models for each segment were developed based on the eye movement data. The models for the segment based on a driver’s cognitive characteristics showed higher accuracies than the models for the male and female groups. This could be a preliminary study introducing the importance of using drivers’ cognitive characteristics in designing visual information for vehicles.