Divergence Between Flight Crew Mental Model and Aircraft System State in Auto-Throttle Mode Confusion Accident and Incident Cases

Abstract

In many aircraft accidents relating to pilot error, flight crews held an inconsistent assumption of the actual state of the system when reacting to a failure or abnormal situation. This paper formally defines divergence between flight crew mental model and actual system state, describes a developed framework to understand the mechanisms behind divergence, and discusses results of analysis of divergence in eight auto-throttle mode confusion accidents and incidents. The framework revealed that divergence occurred in all auto-throttle cases. As divergence was analyzed, it became apparent that reconvergence was also prominent in the accidents and incidents, where seven of eight cases indicated that the crew's mental model and actual system state regained consistency at some point during the accident timeline. Success of recovery actions was dependent on the timing of reconvergence in relation to the criticality of the situation. Consistent with existing literature, the framework also revealed that breakdowns in situation awareness were apparent during the onset of divergence and as an inhibition of reconvergence. The analysis using the framework of divergence showed that for auto-throttle mode confusion accidents, mitigation efforts would be most effective if targeted toward promoting situation awareness of system state prior to the situation degrading beyond recoverability.

A New Interface to Cope With Unreliable Airspeed Indications

“…Further research is also needed on the way the Sycopaero interface should be displayed and reversed in the event of, respectively, an airspeed failure and its resolution. We think that the automatic display of the Sycopaero interface when the system detects airspeed inconsistencies (as it was implemented in the present study) would make these failures – which can sometimes be difficult to detect – more explicit for the pilots and strongly facilitate the detection of the system change of state (Silva & Hansman, 2015). It may also be helpful to allow pilots to manually activate the interface in the case of doubt and/or to decide whether they want to cross out the airspeed indicator or not.…”

Shared control versus traded control in driving: a debate around automation pitfalls

“…Mode errors have been implicated in aviation accidents (Silva and Hansman 2015), oil spill accidents (Seastreak Wall Street; National Transportation Safety Board, 2014) 2 , and industrial accidents (Three Mile Island; Rogovin and Frampton 1980), and also start to be of concern in daily tasks in which automation finds its entrance (Andre and Degani 1997), such as driving (Wilson et al 2020). Current automated driving Figure 2.…”

Situation Awareness Adaptive Alerting in an Aircraft Cockpit: A Simulator Study

“…During monitoring, the pilot continuously compares their expectation about the aircraft's state to the indications by cockpit instruments. When the pilot fails to efficiently monitor the relevant aspects of their environment, their mental picture of the situation can divert from the actual state resulting in poor decision making (Silva & Hansman, 2015).…”