Return to Manual Control After Monitoring Automated Systems: Effects of Different Levels of Reliability

Abstract

Automated systems changed the role of the human operator from active controller to supervisor of automated system. This new role comes with several human factors issues, as monitoring inefficiency. Some authors showed that it is possible to mitigate this monitoring performance drop using adaptive automation. Here, we assessed the consequences of a return to manual control with four different levels of automation reliability on eighty students. We compared the effect of the introduction of two manual sessions to exclusively automated sessions (total duration 90 min) using the MATB-II (i.e., intermittent vs permanent automation). Results showed that participants spend more time looking at the automated task in the 0% level reliability compared to 100% level of reliability over the time whatever the automation type. Manuals sessions in intermittent automation group had an impact on the time spend looking at the automated monitoring task.