FOCUS: An Intelligent Startle Management Assistant for

Maximizing Pilot Resilience

Alexandre Duchevet

, Dong-Bach Vo

, Vincent Peyruqueou,

Théo De-La-Hogue, Jérémie Garcia

, Mickaël Causse

and Jean-Paul Imbert

Fédération ENAC ISAE-SUPAERO ONERA, Université de Toulouse, France

Keywords: Smart Assistant, Startle Effect, Surprise, Human-Machine Teaming, Unexpected Events, Biofeedback,

Situation Awareness.

Abstract: On the flight deck, the startle effect is triggered by sudden, unexpected and possibly threatening events such

as bird strikes or system failures. It is a very rapid protective and defensive reaction that can lead to a partial

or total incapacitation of pilots with tragic consequences. With Single Pilot and Reduced Crew operations

likely being implemented in the future, the single pilot will be forced to face and handle the startle effect alone

in the cockpit. We designed and developed FOCUS (Flight Operational Companion for Unexpected

Situations), an intelligent assistant designed to help single pilots overcome the startle effect. FOCUS supports

pilots in regulating their stress and maintaining an adequate situational awareness. The assistant guides them

to breathe at a specific pace thanks to ambient light brightness variation of the cockpit and draws their

attention towards potentially unseen information. To test and improve its design, we evaluated FOCUS in an

A320 simulator with five qualified pilots. In these trials, FOCUS was positively welcomed by pilots as they

perceived it as a valuable addition to the cockpit. These evaluations will guide further iterations of FOCUS

design and support the understanding of human-AI teaming in the cockpit in subsequent studies.

1 INTRODUCTION

In aviation, the term "startle effect" has been often

used to designate that pilots were literally "startled"

but also "surprised" upon an unexpected event

occurrence. While startle designates the physiological

reflex following an abrupt and intense stimulus

(Koch, 1999), surprise is an emotion resulting from

the discrepancy between one's perception and

expectation. Surprise promotes “cognitive and

motivational processes directed at a proper

understanding of the unexpected event” (Horstmann,

2006). In the cockpit, a startle is almost always

accompanied by surprise (e.g TCAS alerts, lightning

strike). Surprise can also occur without startle and is

primarily due to an unexpected aircraft position, ATC

clearances, or other crewmember actions (Kochan et

https://orcid.org/0000-0002-9751-1194

https://orcid.org/0000-0002-2391-5583

https://orcid.org/0000-0001-7076-6229

https://orcid.org/0000-0002-0601-2518

https://orcid.org/0000-0001-5082-1374

al., 2004). With the growing complexity of cockpits,

automatisms are also a common source of surprise

(Dehais et al., 2015). Startle and surprise can affect

the motor (May & Rice, 1971; Vlasak, 1969) and

cognitive (Thackray et al., 1983; Woodhead, 1958)

capabilities of individuals. On the flight deck, it can

lead to task interruptions, difficulties for reframing,

or inappropriate cockpit inputs (BEA, 2012; KNKT,

2014).

Pilots’ recurrent training and, more generally,

experience are the first barriers against the startle

effect. In simulators, pilots become indeed familiar

with many situations, preparing them to react

accordingly in operations. Crew Resource

Management (CRM) helps also in the startle effect

consequences mitigation. It can be characterized by

several crew competencies like leadership, teamwork,

Duchevet, A., Vo, D., Peyruqueou, V., De-La-Hogue, T., Garcia, J., Causse, M. and Imbert, J.

FOCUS: An Intelligent Startle Management Assistant for Maximizing Pilot Resilience.

DOI: 10.5220/0012915600004562

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Conference on Cognitive Aircraft Systems (ICCAS 2024), pages 5-12

ISBN: 978-989-758-724-5

communication and decision making (ICAO, 2013). In

2016, Field et al. demonstrated that crews that handled

unexpected events well were the ones that had good

CRM, while poorly performing crews showed

mediocre CRM. All the procedures already in place

can also help anticipate unexpected situations and

provide a framework of action to react without

engaging extensive cognitive resources. Finally, the

cockpit design plays a major role in minimizing the

consequences of the startle effect. Recent studies tested

new procedures to counter the startle effect. The

procedures URP (EASA, 2015) and COOL (Landman

et al., 2020) invite pilots to try to relax, breathe before

acting precipitately, and then promote launching the

cognitive process by focusing on what is perceived by

the pilots before making decisions.

All these mitigation measures have helped

prevent many crashes in the past, such as US Airways

1549 accident (NTSB, 2010). In this case, excellent

CRM upon a bird strike saved the situation. However,

these measures suffer from limitations. First, pilots

cannot be trained to face all possible situations, and

experiencing a situation in a simulator is not the same

as in real life (Casner et al., 2013). Moreover, it is

difficult to trigger startle and surprise in simulators.

To this end, Burki-Cohen (2010) recommend an

extremely realistic environment like Full Flight

Simulators to maximize the chances of startle and

surprise. However, as stated in the AF447 report

(BEA, 2012), "Initial and recurrent training as

delivered today does not promote and test the

capacity to react to the unexpected." Finally, even if

procedures are put in place to prevent and minimize

the startle effect, it is very difficult for pilots to apply

them correctly (Grant et al., 2018; Schroeder et al.,

2014). In the future, Single Pilot and Reduced Crew

Operations are likely to be implemented. While the

ambition of this type of operations is to at least keep

the same level of safety, the impossibility to rely on a

second pilot to face the startle effect is a danger that

cannot be overlook.

With the advance in artificial intelligence, new

systems could be implemented to support single pilots

in various tasks and situations, particularly upon

unexpected events occurrence. In 1997 already,

Strohal & Onken presented the concept of CAMA, a

cockpit assistant for military crew. Focusing on

Human-centered automation, one of its functions was

to recognize pilot intents and errors. With the same

objective of understanding crew mental state, the

Crew Workload Manager of Dorneich et al. (2011)

aims to measure objectively the workload of each

crew member to warn about potential workload

imbalance in civil aircraft cockpits. Klaproth et al.

(2020), tried also to understand pilots’ cognitive state

thanks to their brain activity immediately following

an auditory event. Finally, Duchevet et al. (2022) and

Bejarano et al. (2022) developed the HARVIS

assistant which, on one hand, supports pilots taking

the go-around decision thanks to machine learning

and on the other hand, helps them choose the most

appropriate airport during rerouting activities.

These advancements in Artificial Intelligence and

the perspectives of future single pilot operations lead

us to ask ourselves if a cockpit assistant could be

designed to support pilots under startle effect. The

study depicted here is an attempt to answer this

question. In this paper, we introduce FOCUS, our

Flight Operational Companion for Unexpected

Situations prototype, designed to help pilots

overcome the effects of startle and surprise in the

cockpit. In a first part, we detail the concept of

FOCUS through its two functions: the stress

regulation support and the situation awareness

support. In a second part, we report the results of the

qualitative evaluations of the prototype, during which

five professional pilots tested the intelligent assistant

in an A320 simulator on two startling and surprising

scenarios.

2 INTELLIGENT ASSISTANT

CONCEPT

FOCUS employs two strategies to assist pilots in

overcoming the effect of startle. First, it helps pilots

to regulate their stress on startle event occurrence.

Directly inspired from the first step of the URP

(EASA, 2015) and COOL (Landman et al., 2020)

procedures, we want the assistant to guide pilots to

actively relax without demanding too much cognitive

resources. Second, as there is a risk of loss of situation

awareness upon unexpected events occurrence, the

assistant will ensure that no important information is

missed by the pilot, allowing to build a good mental

representation of the situation. The assistant is

supposed to have a startle effect detection module, but

this function has not been developed in the current

state of the prototype. The automatic detection of

startle and surprise would enable a dynamic task

allocation between the pilot and the assistant, the later

ensuring the flight parameter monitoring role.

2.1 Stress Regulation Function

The assistant helps limit the negative effects of acute

stress following two methods. The first method uses

ICCAS 2024 - International Conference on Cognitive Aircraft Systems

a green halo visual effect on cockpit screens and a

green ambient light in the cockpit to guide pilots

breathing deeply. As the light intensity on the screens

and in the cockpit increases, the pilot breathes in, as

the light intensity decreases, the pilot breathes out

(Figure 1). This procedure is based on the Heart-

Focus Breathing technique that is a common step to

increase cardiac coherence where each breathing

cycle last for 10 seconds (McCraty & Zayas, 2014).

Figure 1: Stress regulation function cockpit integration.

The effect of deep breathing on stress has been

extensively demonstrated (Perciavalle et al., 2017).

Breathing can be slowed down to a constant rate to

increase relaxation and reducing stress levels

(Schlatter et al., 2022). However, reaching an optimal

breathing pace requires self-awareness and control.

Therefore, the assistant will help pilots to reach

cardiac coherence, a state where the heart rate

variability is better controlled, to reduce the effect of

stress.

The second method to limit the effect of stress is

to provide pilots with vibrotactile feedback. The pilot

is equipped with a device on the wrist, which provides

a simulated heartbeat through vibration every second,

i.e. 60 beats per minutes, when the stress regulation

support is active. It is worth noting that any wrist-

worn vibrating device such as now common smart

watches could be used for this type of stress

regulation intervention. Research has indeed shown

that constant low heart rate feedback can significantly

reduce anxiety (Sun et al., 2023). Experiencing tactile

feedback with a simulated heartbeat at 60 beats per

minute can reduce perceived anxiety (Costa et al.,

2016) and heart rate after some physical efforts (Choi

& Ishii, 2020).

2.2 Situation Awareness Enhancement

Function

Upon an unexpected event, an orienting response is

elicited, drawing the attention towards the stimulus

(Bradley, 2009). In addition, the task may be

interrupted (Altmann & Trafton, 2004). As situations

can be highly dynamic in the cockpit, a novel event

can therefore lead to a loss of global situation

awareness. To counter this, FOCUS supports pilots in

their perception of the elements in the environment

(SA level 1) (Endsley, 1995). When the support is

active, the assistant highlights the flight parameters

that need to be checked such as altitude or speed.

When highlighted, the parameter is wrapped with a

coloured rectangular box (Figure 2). As the pilot

glances at the instrument, the highlight disappears. It

is made possible by an eye tracking system that track

pilot’s gaze position.

There are two levels of alerts based on an

estimated level of criticality. The first level is

"Caution," and the second level is "Warning". These

levels are represented by two distinct designs on the

Primary Flight Display (PFD), Navigation Display

(ND), and Engine/Warning Display (ECAM). A

situation awareness score is calculated for each

parameter as a function of time, importance of the

parameter, and speed of parameter change. As the

pilot does not look at a parameter, the score

associated to it will start dropping. If the situation

awareness score drops below a certain threshold, the

first level of alert is activated to draw the attention of

the pilot. If the drop continues, the second level is

activated in following.

Figure 2: Situation awareness support and alert levels.

2.3 Assistant’s Human-Machine

Interface

On assistant start, a subliminal icon representing two

humans supporting each other appears on the PFD,

ND and ECAM displays for 500ms to make pilots

aware of the support activation and raise the level of

anthropomorphism of the assistant (de Visser et al.,

2017) (Figure 3).

FOCUS: An Intelligent Startle Management Assistant for Maximizing Pilot Resilience

Figure 3: Assistant's icon.

The pilot can control the assistant through the

Electronic Flight Bag placed on the window side of

the piloting position. Each support function can be

activated and disabled manually on demand, giving

the pilot full control of the assistant. The pilot keeps

this way the complete responsibility of the flight. The

monitor also displays information which helps pilots

always understand the operation of the intelligent

assistant (Figure 4).

The pilot can monitor his heart rate through the

assistant. This piece of information is useful to raise

one’s self-awareness of his own physiological state

and to create an opportunity to enable stress

regulation support if needed. It also allows the

assistant to provide information about its functioning

and an explanation for self-activation to the pilot.

The pilot can monitor the relevancy of his

situation awareness. A global situation awareness

score shown on the display indicates whether flight

parameters reading is required. In addition, a widget

mimicking the PFD suggests which areas of interest

should be checked to increase the global situation

awareness score. This contributes to assistant’s

transparency and provide opportunities to improve

the pilot’s situation awareness.

The interface enables pilots to cancel the

automatic support execution when the intelligent

assistant has identified a situation of startle. This

ensures that when the assistant support is not

required, the system will not interfere with the

piloting tasks.

Figure 4: Assistant’s HMI: stress regulation, situation

awareness and auto-support monitor and control.

3 METHOD

An evaluation of the assistant prototype was

conducted to collect professional pilots’ feedback on

the design in a realistic environment. The stress

regulation and the situation awareness supports were

implemented in the prototype. The startle effect

detection was played in wizard of Oz.

3.1 Participants

Five pilots participated to this preliminary study (All

male, mean age=43.8, SD=7.5). Four were qualified

on A320 and one was qualified on Boeing aircraft

(mean number of flights hours=4600, SD=5672). The

protocol was approved by the ethics commission for

research of the university of Toulouse (project 2023-

740).

3.2 Scenarios

Two scenarios were created to evaluate the assistant:

1) a lightning strike on final approach, aiming to

generate startle and surprise, 2) a shifting cargo at

take-off aiming to create a surprise.

In the lightning strike scenario, the aircraft is

struck by lightning on final approach. As a result, a

loud bang is heard, and an intense flash is triggered,

provoking startle and surprise. Because of the

lightning strike, electrical problems on board of the

aircraft lead to the disconnection of automatisms.

The cargo shift scenario occurs shortly after take-

off. A cargo gets loose and a shift of centre of gravity

occurs. As a result, a strong pitch up moment is

observed, triggering a surprise. The pilot is forced to

react quickly to control the aircraft and a rapid

landing is necessary.

3.3 Physiological Measures

To assess if participants were startled and surprised,

they were equipped with physiological sensors to

monitor the cardiac (PPG, ECG), electrodermal

(GSR) and muscular activity (EMG). Participants’

reactions and facial expressions were filmed for

ground truth.

3.4 Procedure

Upon consent form completion, each participant was

first introduced to the goal and the proceedings of the

study, the context and the intelligent assistant

functions. The two flight scenarios of the study were

then presented without disclosing the startling and

ICCAS 2024 - International Conference on Cognitive Aircraft Systems

surprising events (lightning strike and cargo shift). In

each scenario, the flight was conducted in Single Pilot

Operations. Each flying session took place in an

Airbus A320 simulator.

Upon eye tracking calibration, one of the

experimenters performed a walkthrough of the

intelligent assistant functions and invited the

participant to experience the stress regulation and the

situation awareness supports. When ready, the

participant began the training phase, which lasted for

about 20min, aiming at familiarizing with the

simulator during a take-off/landing scenario. In

addition, the participant was allowed and encouraged

to request support and to experience the assistant

functions during the training.

After completion of the training, the participant

performed the two validation scenarios in random

order. At the end of the scenarios, the participant was

debriefed about the experience and performance

during the flight. Between each scenario, the

participant filled a Likert-scale questionnaire to

assess the subjective perception on the performance,

the usefulness and the understanding of the intelligent

assistant. Finally, when the two scenarios were

completed, the participant was invited to debrief

about the intelligent assistant support through a semi-

structured interview in which usability,

improvements, AI initiative and trust was discussed.

All pilots performed the lightning strike scenario.

Because of a lack of time during some evaluation

sessions, only three of them performed the cargo-shift

scenario in addition.

4 RESULTS

Although it is not possible to report statistics analysis

given the sample size of the study, the questionnaire

results and the semi-structure interviews analysis can

provide experienced pilots’ qualitative feedback on

FOCUS.

4.1 General Feedback

The scenarios successfully triggered startle and

surprise in the participants. On a scale from 1 (not

startled or surprised) to 10 (very startled or surprised),

participants reported an average score of 7.0

(SD=3.06) for startle and an average score of 7.6

(SD=1.51) for surprise in the lightning strike

scenario. The cargo shift scenario was deemed less

intense with a startle average score of 3.7 (SD=2.52)

and a surprise average score of 5 (SD=2.64).

Physiological data (Figure 5) and facial expressions

confirmed that all the participants were startled or

surprised during the lightning strike scenario. Signs

of stress after the surprising event were observed in

physiological data in the cargo shift scenario for all

the participants. For example, P4 commented about

the lightning strike scenario: “I was sort of surprised

and trying to figure out what is working and where I

am, what direction am I going?”.

The assistant was generally welcomed by the

participants. All of them were aware of the assistant’s

activation thanks to the icon appearing on screens. P4

stated that the icon disturbed his visual scan but

allowed him to “take a step back”. The participants

thought that the assistant made them able to maintain

a good situation awareness and that the awareness

guidance was overall relevant. The system actions

and purpose were well understood by the participants,

and the assistant was thought easy to interact with.

Participants felt somewhat confident to work with the

assistant. That being said, the participants felt unsure

about the benefits of the assistant to limit the

detrimental effect of startle and surprise, and its

usefulness when unexpected events occurred (i.e. in

situations of surprise).

Figure 5: Example of a participant's electrodermal activity

(Galvanic Skin Response) following the lightning strike

event, resulting in the increase of the conductance.

4.2 Stress Regulation Support

The breathing guidance lights were seen as an

indicator of stress level during the exercise (P1). By

seeing the light, P4 was reminded to breathe deeply.

However, participants highlighted the lack of

availability to perform the breathing procedure. For

instance, P2 stated: “You need availability to focus on

your cardiac rhythm” and P3 added: “I was thinking

about it during the last [led] cycle, then I tried to

adapt my breathing. Before that, I don’t think I was

aware of or adapting my breathing”. Some pilots

admitted that focusing on their breathing while

dealing with the aircraft automation failures was quite

difficult.

FOCUS: An Intelligent Startle Management Assistant for Maximizing Pilot Resilience

The tactile heartbeat simulation was not found

uncomfortable. No participant noticed the tactile

feedback during the exercise even though they

experienced it during the training phase. Pilots were

eager to know whether the tactile feedback had a real

impact on their heartbeat. P2 thought that pilots could

benefit from tactile feedback as it was “transparent”

(i.e. unintrusive) to him and to other pilots, and that it

did not have any impact on his workload. He added

that the technology may be promising if it can be

compatible with existing pilot smartwatches.

Finally, it is worth noting that none of the

participants found the physiological monitoring

display useful. Even though P3 was monitoring his

physiological status at the beginning of the exercise,

he stopped when the assistant support started. The

other pilots reported that they did not look at the

assistant control interface on the Electronic Flight

Bag during the exercises. Pilots reported indeed that

they did not have the cognitive resources to check the

physiological monitoring display once the emergency

declared. They had to focus on handling the

emergency and controlling the aircraft.

4.3 Situation Awareness Support

The situation awareness support was useful to

participants, especially when the autopilot failed in

our exercise (M=3.85/5, SD=0.69). The assistant

managed to draw pilots’ attention towards specific

parameters. P5 commented: “I missed the speed

change. I was glad that the assistant told me to look

at the speed”. P4 thought the benefit of the situation

awareness support was to “increase the sampling

rate” to acquire flight information. He thought that it

may have focused too long on the navigation display

or on some other information. P3 said that the

attention getter helped him to check the right pieces

of information, even though this is what he was

already planning. He reported: “I think I would have

done it but it allowed me to save time”. P1 also

commented on the potential for such assistance to

support pilot flying aircraft that they are not familiar

with. With more training with the assistant, the

participants thought they could follow its guidance

better.

However, they also warned about the potential

distraction that could result from the PFD red

highlighting boxes (Warning level). Because some

pilots did not look to all the instruments, more boxes

started to appear on the PFD. They thought that this

was overwhelming and confused them about what

instrument to look first. The pilot’s inability to make

some boxes disappear upon glancing at it was partly

due to an assistant lack of performances in

recognizing Areas of Interest (AOI) in the cockpit.

Particularly, it failed several times to detect the AOI

associated with the heading and the localizer

deviation on the Primary Flight Display resulting in

constant red boxes around these zones even though

pilots were looking at them.

5 DISCUSSION

Our study showed a successful cockpit integration of

an intelligent assistant to support pilots during startle

effect in Single Pilot Operations. As shown by

participants’ statements, the need to maintain a good

situation awareness upon unexpected events

occurrence appears to be strong and a stress

regulation system seems promising. However, limits

to our approach should be mentioned. As no

commercial airliner is today flown single pilot, the

most adequate solution for FOCUS’ evaluation was

to perform simulations in an Airbus A320 cockpit,

even though it is not designed specifically for regular

Single Pilot Operations. Moreover, the training to use

the assistant was short (20 minutes) compared to

standard training for new systems. With more

experience with FOCUS, pilots could be more at ease

with interacting with the assistant which could bring

new perspectives and feedback.

Evaluating FOCUS highlighted key points for

designing assistants in highly dynamic and complex

situations. Even if agent transparency is vital,

providing information to the pilots should not

overload and disturb her/him. Thus, the availability of

cognitive resources to process information appears to

be one of the main challenges for startle effect

management. We believe that one way to pursue our

research would be to look for the most appropriate

means to reduce stress in dynamic and cognitively

demanding environments. Furthermore, as FOCUS

adapts to pilots’ behaviour thanks to the analysis of

the gaze position, we could improve the situation

awareness support by pushing the adaptiveness of the

assistant to fit exactly pilots’ profile and own scan

path.

FOCUS is an attempt to design and develop a

Human-Automation Teaming (HAT) agent for the

cockpit. It aligns with the conceptual model of

Shively et al. (2018). First, FOCUS achieves dynamic

task allocation by assuming a monitoring role when a

startle effect is detected. Second, bidirectional

communication is present, with the pilot passively

sending gaze information and physiological data,

while the assistant suggests where to focus attention.

ICCAS 2024 - International Conference on Cognitive Aircraft Systems

This passive communication from the pilot to the

Agent might be a solution to improve team situation

awareness and to communicate effectively in a

dynamic situation (Demir et al., 2017), with pilots

constantly pushing information without effort to the

HAT agent. Finally, transparency, as with other

cockpit systems, poses a challenge for FOCUS.

Pilots, especially under the startle effect, have limited

cognitive and time resources to comprehend the

reasoning behind FOCUS outputs. Its transparency

and trust in it may develop through training and post-

operations analysis of agent outputs, rather than

solely during operations, echoing directly a research

gap depicted by Lyons et al. (2021): "What is the

optimal level and method of training to team with

machines [...]?”. But building a solid trust in the HAT

agent during training and rely exclusively on it when

things go south seems dangerous. The challenge lies

therefore is the following question: How to design a

transparent artificial teammate while the cognitive

capacities of the human counterpart are temporary

diminished?

ACKNOWLEDGEMENTS

This work was done as part of the HAIKU project.

This project has received funding from the European

Union’s Horizon Europe research and innovation

programme HORIZON-CL5-2021-D6-01-13 under

Grant Agreement no 101075332.

The authors wish to acknowledge all the

professional pilots that took their time to share their

useful insight during interviews, design meetings and

simulations. A special thanks to Yves Rouillard,

flight simulator manager, without whom the

evaluations of the assistant would not have been

possible.

REFERENCES

Altmann, E. M., & Trafton, J. G. (2004). Task Interruption :

Resumption Lag and the Role of Cues. Proceedings of

the Annual Meeting of the Cognitive Science Society,

26(26).

BEA. (2012). Rapport final, accident survenu le 1er Juin

2009, vol AF447 Rio de Janeiro—Paris. https://bea.

aero/docspa/2009/f-cp090601/pdf/f-cp090601.pdf

Bejarano, C., Vázquez, A. L. R., Colomer, A., Cantero, J.,

Ferreira, A., Moens, L., Duchevet, A., Imbert, J.-P., &

Hogue, T. D. L. (2022). HARVIS : Dynamic rerouting

assistant using deep learning techniques for Single Pilot

Operations (SPO). Transportation Research Procedia,

66, 262‑269. https://doi.org/10.1016/j.trpro.2022.12.0

Bradley, M. M. (2009). Natural selective attention :

Orienting and emotion. Psychophysiology, 46(1), 1‑11.

https://doi.org/10.1111/j.1469-8986.2008.00702.x

Burki-Cohen, J. (2010). Technical Challenges of Upset

Recovery Training : Simulating the Element of

Surprise. In AIAA Modeling and Simulation

Technologies Conference. American Institute of

Aeronautics and Astronautics. https://doi.org/10.2514/

6.2010-8008

Casner, S. M., Geven, R. W., & Williams, K. T. (2013). The

Effectiveness of Airline Pilot Training for Abnormal

Events. Human Factors, 55(3), 477‑485.

https://doi.org/10.1177/0018720812466893

Choi, K. Y., & Ishii, H. (2020). ambienBeat : Wrist-worn

Mobile Tactile Biofeedback for Heart Rate Rhythmic

Regulation. Proceedings of the Fourteenth

International Conference on Tangible, Embedded, and

Embodied Interaction, 17‑30. https://doi.org/10.1145/

3374920.3374938

Costa, J., Adams, A. T., Jung, M. F., Guimbretière, F., &

Choudhury, T. (2016). EmotionCheck : Leveraging

bodily signals and false feedback to regulate our

emotions. Proceedings of the 2016 ACM International

Joint Conference on Pervasive and Ubiquitous

Computing, 758‑769. https://doi.org/10.1145/2971648.

2971752

Dehais, F., Peysakhovich, V., Scannella, S., Fongue, J., &

Gateau, T. (2015). « Automation Surprise » in

Aviation : Real-Time Solutions. Proceedings of the

33rd Annual ACM Conference on Human Factors in

Computing Systems, 2525‑2534. https://doi.org/10.11

45/2702123.2702521

Demir, M., McNeese, N. J., & Cooke, N. J. (2017). Team

situation awareness within the context of human-

autonomy teaming. Cognitive Systems Research, 46,

3‑12. https://doi.org/10.1016/j.cogsys.2016.11.003

de Visser, E. J., Monfort, S. S., Goodyear, K., Lu, L.,

O’Hara, M., Lee, M. R., Parasuraman, R., & Krueger,

F. (2017). A Little Anthropomorphism Goes a Long

Way : Effects of Oxytocin on Trust, Compliance, and

Team Performance with Automated Agents. Human

Factors, 59(1), 116‑133. https://doi.org/10.1177/00187

20816687205

Dorneich, M. C., Passinger, B., Hamblin, C., Keinrath, C.,

Vasek, J., Whitlow, S. D., & Beekhuyzen, M. (2011).

The Crew Workload Manager : An Open-loop Adaptive

System Design for Next Generation Flight Decks.

Proceedings of the Human Factors and Ergonomics

Society Annual Meeting, 55(1), 16‑20.

https://doi.org/10.1177/1071181311551004

Duchevet, A., Imbert, J.-P., Hogue, T. D. L., Ferreira, A.,

Moens, L., Colomer, A., Cantero, J., Bejarano, C., &

Vázquez, A. L. R. (2022). HARVIS : A digital assistant

based on cognitive computing for non-stabilized

approaches in Single Pilot Operations. Transportation

Research Procedia, 66, 253‑261. https://doi.org/

10.1016/j.trpro.2022.12.025

FOCUS: An Intelligent Startle Management Assistant for Maximizing Pilot Resilience

EASA. (2015). EASA_Research_Startle_Effect_Manage

ments_Final_Report.pdf.

https://www.easa.europa.eu/sites/default/files/dfu/EAS

A_Research_Startle_Effect_Managements_Final_Rep

ort.pdf

Endsley, M. R. (1995). Toward a Theory of Situation

Awareness in Dynamic Systems. Human Factors,

37(1), 32‑64. https://doi.org/10.1518/0018720957790

49543

Field, J., Fucke, L., Correia Grácio, B., & Mohrmann, F.

(2016, juin 13). Flight Crew Response to Unexpected

Events : A Simulator Experiment. AIAA Modeling and

Simulation Technologies Conference. AIAA Modeling

and Simulation Technologies Conference, Washington,

D.C. https://doi.org/10.2514/6.2016-3373

Grant, P. R., Moszczynski, G. J., & Schroeder, J. A. (2018,

juin 25). Post-stall Flight Model Fidelity Effects on Full

Stall Recovery Training. 2018 Modeling and

Simulation Technologies Conference. 2018 Modeling

and Simulation Technologies Conference, Atlanta,

Georgia. https://doi.org/10.2514/6.2018-2937

Horstmann, G. (2006). Latency and duration of the action

interruption in surprise. Cognition & Emotion, 20(2),

242‑273. https://doi.org/10.1080/02699930500262878

ICAO. (2013). International Civil Aviation Organisation

[ICAO]. “Manual of Evidence-Based Training ”, (Doc

9995). Montreal, QC: ICAO.

Klaproth, O. W., Vernaleken, C., Krol, L. R., Halbruegge,

M., Zander, T. O., & Russwinkel, N. (2020). Tracing

Pilots’ Situation Assessment by Neuroadaptive

Cognitive Modeling. Frontiers in Neuroscience, 14.

https://www.frontiersin.org/articles/10.3389/fnins.202

0.00795

KNKT. (2014). Komite Nasional Keselamatan

Transportasi (2014), Aircraft Accident Investigation

Report, Indonesia Air Asia Airbus A320-216; PK-AXC.

Koch, M. (1999). The neurobiology of startle. Progress in

Neurobiology, 59(2), 107‑128. https://doi.org/10.1016/

S0301-0082(98)00098-7

Kochan, J., Breiter, E., & Jentsch, F. (2004). Surprise and

Unexpectedness in Flying : Database Reviews and

Analyses. Th ANNUAL MEETING, 5.

Landman, A., van Middelaar, S. H., Groen, E. L.,

van Paassen, M. M. (René), Bronkhorst, A. W., &

Mulder, M. (2020). The Effectiveness of a Mnemonic-

Type Startle and Surprise Management Procedure for

Pilots. The International Journal of Aerospace

Psychology, 30(3‑4), 104‑118. https://doi.org/10.1080/

24721840.2020.1763798

Lyons, J. B., Sycara, K., Lewis, M., & Capiola, A. (2021).

Human–Autonomy Teaming : Definitions, Debates,

and Directions. Frontiers in Psychology, 12, 589585.

https://doi.org/10.3389/fpsyg.2021.589585

May, D. N., & Rice, C. G. (1971). Effects of startle due to

pistol shots on control precision performance. Journal

of Sound and Vibration, 15(2), 197‑202.

https://doi.org/10.1016/0022-460X(71)90534-7

McCraty, R., & Zayas, M. A. (2014). Cardiac coherence,

self-regulation, autonomic stability, and psychosocial

well-being. Frontiers in Psychology, 5.

https://doi.org/10.3389/fpsyg.2014.01090

NTSB. (2010). Loss of Control on Approach Colgan Air,

Inc. Operating as Continental Connection Flight 3407

(Accident Report NTSB/AAR-10/01 PB2010-910401;

p. 299). National Transportation Safety Board.

https://www.ntsb.gov/investigations/accidentreports/re

ports/aar1001.pdf

Perciavalle, V., Blandini, M., Fecarotta, P., Buscemi, A., Di

Corrado, D., Bertolo, L., Fichera, F., & Coco, M.

(2017). The role of deep breathing on stress.

Neurological Sciences, 38(3), 451‑458. https://doi.org/

10.1007/s10072-016-2790-8

Schlatter, S. T., Thérond, C. C., Guillot, A., Louisy, S. P.,

Duclos, A., Lehot, J.-J., Rimmelé, T., Debarnot, U. S.,

& Lilot, M. E. (2022). Effects of relaxing breathing

paired with cardiac biofeedback on performance and

relaxation during critical simulated situations : A

prospective randomized controlled trial. BMC Medical

Education, 22(1), 422. https://doi.org/10.1186/s12909-

022-03420-9

Schroeder, J. A., Bürki-cohen, J., Shikany, D. A., Gingras,

D. R., & Desrochers, P. (2014). An Evaluation of

Several Stall Models for Commercial Transport

Training.

Shively, R. J., Lachter, J., Brandt, S. L., Matessa, M.,

Battiste, V., & Johnson, W. W. (2018). Why Human-

Autonomy Teaming? In C. Baldwin (Éd.), Advances in

Neuroergonomics and Cognitive Engineering (Vol.

586, p. 3‑11). Springer International Publishing.

https://doi.org/10.1007/978-3-319-60642-2_1

Strohal, M., & Onken, R. (1998). Intent and error

recognition as part of a knowledge-based cockpit

assistant. Applications and Science of Computational

Intelligence, 3390, 287‑299. https://doi.org/10.1117/

12.304818

Sun, Y., Lu, T., Wang, X., Chen, W., Chen, S., Chen, H., &

Zheng, J. (2023). Physiological feedback technology

for real-time emotion regulation : A systematic review.

Frontiers in Psychology, 14, 1182667. https://doi.org/

10.3389/fpsyg.2023.1182667

Thackray, R. I., Touchstone, R. M., & Civil Aeromedical

Institute. (1983). Rate of initial recovery and

subsequent radar monitoring performance following a

simulated emergency involving startle. (FAA-AM-83-

13). https://rosap.ntl.bts.gov/view/dot/21242

Vlasak, M. (1969). Effect of startle stimuli on performance.

Aerospace Medicine, 40(2), 124‑128.

Woodhead, M. M. (1958). The Effects of Bursts of Loud

Noise on a Continuous Visual Task. British Journal of

Industrial Medicine, 15(2), 120‑125.

ICCAS 2024 - International Conference on Cognitive Aircraft Systems