Predicting human detection of automation failures: towards humanautomation integration

Natalie Griffiths

Research output: ThesisDoctoral Thesis

65 Downloads (Pure)

Abstract

This thesis explored methods for proactively predicting operators’ RTM performanceprior to automation failing using simulated ATC as a complex task environment testbed. Studies 1and 2 examined operator states (workload, fatigue, trust in automation, and task engagement) aspredictors of subsequent RTM performance following an automation failure under high or low degreeof automation (DOA). The findings provided preliminary support for the prediction of subsequent RTMperformance from dynamic changes in operator states. Study 3 furthered this predictive approach byexamining the relationship between manual skill and automation failure detection performance in highand low DOA.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • The University of Western Australia
Supervisors/Advisors
  • Bowden, Vanessa, Supervisor
  • Loft, Shayne, Supervisor
  • Wee, Serena, Supervisor
Thesis sponsors
Award date13 May 2024
DOIs
Publication statusUnpublished - 2024

Cite this