Driver mental workload and its assessment are becoming more and more important with increased penetration of `Road Transport Informatics' and with changes in driver population and capability, i.e. the increasing number of elderly drivers. In particular newly developed electronic driver support systems, such as route guidance systems, are being tested now for their potential negative effect on driving performance because of the additional mental workload they impose. The increased attention for mental workload techniques has made obvious that there are many factors that complicate accurate assessment of workload in the field. Firstly, there are different causes for increased workload. Both an impaired driver state, e.g. as a result of the use of alcohol, and increased complexity, e.g. an additional task that has to be performed, lead to elevated driver mental workload. Secondly, the measurement techniques themselves are differentially sensitive to changes in workload. Thirdly, the driving task is to a large extent a self-paced task. This means that the driving speed chosen or the accuracy in lane-keeping are adapted by the driver, not only on the basis of external demands but are also dependent upon strategy and self-set goals. Not only these factors in isolation, but also their interaction complicate the measurement of drivers' mental workload. In an effort to understand this, mental workload, task demands and performance were related to each other in a simple model, which was presented. Several of the problems that are encountered when trying to measure mental workload are put in a different perspective in this model.
Abstract