Towards an On-line Assessment of Subjective Driver Workload

The number of Advanced Driving Assistance Systems (ADAS) integrated in modern cars has grown over the last few years (e.g., BAS, ACC, navigation systems). Because car manufacturers must ensure that all these systems are compatible with the driving task, evaluation using an experimental protocol is often necessary. When analyzing the data from such evaluations, the driving context must be accurately taken into account since it has an important influence on the driving task. Thus, methods for assessing “mental workload” are often considered useful tools, as they can take both the driving context (e.g., kind of road, traffic, meteorological conditions) and the use of ADAS into account. Though subjective workload assessment methods have proved quite sensitive in many studies, they are not well suited to on-line assessment. To solve this problem, we have adapted the Instantaneous Self Assessment method (ISA), initially developed for air traffic control, for use during the driving task. An experiment was conducted on the SHERPA full-size driving simulator. Fourteen subjects were asked to complete the same 30-kilometer run twice, once with an Adaptive Cruise Control (ACC; a kind of ADAS) and once without. The route comprised stretches on motorways, major roads and mountain roads, and included various traffic densities. The ISA scale was quite senstive to traffic conditions. Moreover, ISA highlighted the prevailing influence of traffic density over road profile. However, the impact of ADAS use could not be ascertained. This study constitutes a first step towards the development of a new on-line workload assessment technique based on an automatic classification system able to take several indicators into account, including vehicle indicators (e.g., speed, brake and clutch pedals) and driver indicators (e.g., eye-gaze position).