This report describes an experiment in which inexperienced and experienced drivers drove an instrumented car in a number of driving situations in real traffic, while performing one of three secondary tasks. The study's major aim was to find determinants of driver workload which may later serve as elements of an in-car information management system. A distinction was made between load on visual and on cognitive processing resources. The driving situation was found to be the major determinant of visual load in both groups of drivers, whereas it determined cognitive load for only inexperienced drivers. Tasks loading cognitive resources can be carried out in all situations by experienced but not by inexperienced drivers. The results support the notion that experienced drivers can be described by multiple processing resources, whereas inexperienced drivers are better described by a single resource model. No effects of traffic density were found on driver workload. Steering wheel action rate and mirror glance frequency were found to be good indicators for load on visual and cognitive processing resources. Since practically no effects of the secondary tasks were found on driving performance measures, display glance frequency and durations may serve as a safety norm when evaluating in-car information systems.
Abstract