Abstract
Human out-of-the-loop (OOTL) performance problems in overseeing automated systems have motivated an interest in the use of intermediate levels of automation and adaptive automation (AA) as methods for improving the efficiency of operators when working with systems. In this paper, we discuss the current state of research into level of automation (LOA) and AA in complex, dynamic control systems. Different levels of automation and taxonomies of LOA, as well as strategies to AA, are identified. Empirical studies independently demonstrating the effectiveness of LOA and AA for combating negative consequences of OOTL performance and improving human functioning are reviewed. On the basis of these studies, the need to assess the combined effectiveness of LOA and AA in a dynamic control task is established. An experiment is presented which addresses this need. Thirty university students performed the control task and a secondary monitoring task with various levels of automation (varying degrees of computer assistance) and AA (varying durations of assistance) being applied to the former. Testing involved five levels of automation allocated over three different cycle times. Results suggest that LOA and AA are not additive in their effect on automated system functioning, however, each affects performance in very different ways. Level of automation was found to have a significant effect on functioning with the dynamic control task, while AA was found to significantly affect performance on the secondary monitoring task indicating an impact of operator workload. The results of this study are directly applicable to automated system design decisions regarding which system functions should be performed by a human operator or computer controller and for how long performance should occur.
Get full access to this article
View all access options for this article.