Skip to Main content Skip to Navigation

Mind wandering dynamic in automated environments and its influence on out-of-the-loop situations

Abstract : Higher levels of automation are progressively integrated in critical environments to satisfy the increasing demand for safer systems. Such philosophy moves operators to a supervisory role, also called out-of-the-loop (OOTL) situations. Unfortunately, OOTL situations also create a new kind of human-machine interaction issues, called OOTL performance problem. The dramatic consequences of OOTL performance problem stress the need to identify which mechanisms could influence their appearance. The emergence of thoughts unrelated to the here and now, labeled mind wandering (MW), could affect operators in OOTL situations through the perceptual decoupling induced. This thesis investigates MW dynamic in OOTL situations and its influence on operators. We firstly reviewed the evidences in the literature underlining a link between OOTL performance problem and MW. We completed theoretical insights by reporting pilots' tendency (collected with a questionnaire) to encounter more problems with autopilots when experiencing more task-unrelated MW. Then, we conducted three experiments in OOTL conditions using an obstacle avoidance task. With non-expert population and sessions longer than 45 minutes, we observed a significant increase of MW in OOTL situations compared to manual conditions, independently of system reliability. MW episodes were also accompanied by a perceptual decoupling from the task induced by task-unrelated MW. This decoupling was visible on reports of mental demand as well as oculometric (pupil size, blinks) and encephalographic (N1 component, alpha activity) signals. Overall, our results demonstrate the possibility to use physiological markers of MW in complex OOTL environments. We discuss new perspectives towards the use of MW markers to characterize the OOTL performance problem. Instead of blindly stopping MW episodes, which could have benefits for operators, future research should focus on designing systems able to cope with MW and identify information needed to facilitate the reentry in the control loop when needed.
Document type :
Complete list of metadatas

Cited literature [558 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Thursday, October 31, 2019 - 4:04:17 PM
Last modification on : Wednesday, March 4, 2020 - 3:53:50 AM
Long-term archiving on: : Saturday, February 1, 2020 - 5:12:48 PM


Version validated by the jury (STAR)


  • HAL Id : tel-02341861, version 1



Jonas Gouraud. Mind wandering dynamic in automated environments and its influence on out-of-the-loop situations. Human health and pathology. Université Paul Sabatier - Toulouse III, 2018. English. ⟨NNT : 2018TOU30269⟩. ⟨tel-02341861⟩



Record views


Files downloads