A multi-level approach to investigate the influence of the driving event on the driver's cognitive state

Young reseachers seminar 2017, Berlin, ALLEMAGNE, 16-/05/2017 - 18/05/2017Nowadays, the development of new technologies in the automotive industry integrates more and more different complex safety systems. Some of these security systems (ADAS) are based on the monitoring of the driver's cognitive state (Wang, Xu, & Gong, 2010). To this aim, these systems take into account some physiological and/or behavioural parameters in order to estimate for example the level of stress/anxiety or the level of mental workload. Nevertheless, most of these systems do not integrate in their functioning the type of driving situation in which the driver is (e.g., overtaking of a truck, sudden occurrence of a pedestrian). Furthermore, the temporal period for which physiological and behavioral data are affected by a given driving situation still needs to be determined. Therefore, a study has been driven on a simulator so as to investigate how 1) driving could impact drivers' internal state; 2) which indicator (physiological and/or behavioural) was the more sensitive to these changes. Participants had to perform three different driving situations (left-turn, overtaking and pedestrian occurrence) within two sessions (baseline and experimental sessions). The traffic conditions were manipulated to increase traffic density and/or time pressure for each situation of the experimental session, when compared to the baseline session. After each driving situation, the driver had to perform a standardized car following task (Brookhuis, Waard, & Mulder, 1994) during five minutes while the driver's cognitive state was analyzed through physiological (EDA and ECG) and behavioural (driving performance parameters). Besides, subjective measures (questionnaires) were performed after each driving session. The results highlighted the relationship between high mental workload and physiological variations after all experimental situations. Nevertheless, behavioural reactions following the unexpected event seemed to differ from those obtained after a situation inducing anxiety, because of a high traffic density. Moreover, the signatures of behavioural and physiological reactions to the experimental situations did not present the same duration. While, physiological data quickly returned to their baseline value, (after one minute following the situation), behavioural variations persisted on a longer period (variations visible during the five minutes following the situation). Despite the fact that these results have to be confirmed by future studies, suggestions can already be made for the conception of the future ADAS centered on the driver's cognitive state