This article was downloaded by: [Nanyang Technological University] On: 28 April 2015, At: 01:02 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Ergonomics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/terg20
The ergonomic evaluation of eye movement and mental workload in aircraft pilots a
a
b
Yasuko Itoh , Yoshio Hayashi , Ippei Tsukui & Susumu Saito
c
a
Department of Administration Engineering , Keio University , 3-14-1, Hiyoshi, Kohoku-ku, Yokohama, 223, Japan. b
Japan Aeromedical Research Center , 2-3-1, Haneda Airport, Ohta-ku, Tokyo, 144, Japan.
c
National Institute of Industrial Health , Nagao 6, Tama-ku, Kawasaki, 214, Japan. Published online: 27 Mar 2007.
To cite this article: Yasuko Itoh , Yoshio Hayashi , Ippei Tsukui & Susumu Saito (1990) The ergonomic evaluation of eye movement and mental workload in aircraft pilots, Ergonomics, 33:6, 719-732, DOI: 10.1080/00140139008927181 To link to this article: http://dx.doi.org/10.1080/00140139008927181
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
ERGONOMICS, 1990, VOL. 33, No.6, 719-733
The ergonomic evaluation of eye movement and mental workload in aircraft pilots YASUKO ITOH and YOSHIO HAYASHI Department of Administration Engineering, Keio University, 3-14-1, Hiyoshi, Kohoku-ku, Yokohama 223, Japan
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
IpPEI TSUKUI Japan Aeromedical Research Center, 2-3-1, Haneda Airport, Ohta-ku, Tokyo 144, Japan SUSUMU SAITO National Institute of Industrial Health, Nagao 6, Tama-ku, Kawasaki 214, Japan Keywords: Eye movement;Visual scanning; Entropy; Integrated display; Heart rate
variability; Mental workload. This paper presents an experiment which examines characteristics of pilots' scanning behaviour when using integrated CRT displays, and the changes in characteristics when pilots face abnormal situations. The subjects were five experienced pilots. They performed two modes of flight tasks, under normal and abnormal situations, in flight simulators with standard settings. The flight simulators were for a Boeing 747-300(B747), which made use of electromechanical displays, and for a Boeing 767 (B767), equipped with integrated CRT displays. The results showed that the B767pilots tended to gaze at the attitude director indicator which was displayed in the integrated CRT display. It was assumed that 'gaze-type scanning' might be one of the characteristics of pilots' scanning behaviour in cockpits which use the integrated display. By employing subjective ratings and heart rate variability to measure mental workload, no differences in mental workload between the B767pilots and the B747pilots were observed. However, in abnormal situations, the changes in scanning pattern for B767pilots were found to be smaller than those of the B747 pilots. It is concluded that the application of integrated displays helps pilots to obtain sufficient information more easily than electromechanical displays do, even under abnormal situations.
1. Introduction The increasing complexity of navigation procedures and more stringent safety standards for civil transport aircraft have led to a proliferation of instruments and information systems to enable pilots to obtain the essential information to perform flying operations safely and satisfactorily. On the other hand, statistics reveal that the factor which most contributes to total loss accidents caused by crew behaviour is inadequate judgement and misoperation (see figure I). Therefore, information display in cockpits to enable the pilot to make rapid and correct decisions is imperative. Some studies focus on the instrument which provides an integrated presentation of the flight situation and which is adaptable to the different phases of flight. In the early 1980s, an introduction ofcathode-ray tube (CRT) flight displays in civil aircraft marked a watershed in the evolution of information displays in cockpits. The use of CRT 0014-0139/90 $3-00 © 1990 Taylor & Francis Ltd.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
720
Figure I.
Y. Itoh
Factors influencing total loss accidents caused by the crew for the passenger jet planes, worldwide (1959-1983); Source: ICAO Bulletin.
(~)
~oeing
m
.J •..• J. _t.".t_
iu::":U:
(C)
~oeing 141-4~U
Figure 2. Examples of cockpit instrument panels.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
Eye movement and mental workload in pilots
721
displays achieved an integration of instruments, so facilitating the more effective utilization of high priority panel space and greater flexibility. Furthermore, timesharing of the high priority panel space is thereby realized, and information which is not relevant in any particular phase of flight is removed from view altogether. As an example of the integrated displays and electromechanical display, figure 2 shows the control panels of the Boeing 747-300 (8747), the Boeing 767 (8767), and the Boeing 747-400(8747-400). While the earlier electromechanical instruments are used in 8747 cockpits, the new technology of integrated instrument systems using CRT displays has been introduced into 8767 and 8747-400 cockpits. The displays in the 8747-400 cockpit are more advanced than those in the 8767 cockpit, in terms of integration of instruments. For example, the Primary Flight Display in the 8747-400 cockpit integrates the attitude director indicator and the horizontal situation indicator with other instruments into a single display. The 8767 cockpit itself is considerably more integrated than the 8747 cockpit. In order to compare the B767 with the B747, one may take the attitude director indicator as an example of integrated devices. Although the Electro Attitude Director Indicator (EADI) in the B767 cockpit looks somewhat similar to the electromechanical display of the altitude director indicator in the B747 cockpit, the EADI has a more advanced function than the latter, because the EADI can display a tracking target which the plane must head towards. Thus, the EADI provides enhanced support to pilots. Both the B767 and the B747 flight simulators were used in this experiment. It is claimed that the operational benefit of integrated displays is a reduction in the scanning area for essential information, so that the pilots can easily understand flight conditions. On the other hand, pilot performance and workload might be changed by the introduction of such integrated displays. In particular, pilots' eye movements might be influenced. Thus, it is important to ascertain whether or not the introduction of the integrated displays is preferable for pilots. This paper examines the differences in scanning behaviour under conditions of high mental workload between pilots using an integrated display, and pilots using the earlier electromechanical display. The experiment was conducted in B767 and 8747 flight simulators, with standard settings. Pilots were exposed to abnormal situations without any notice, and the changes of scanning behaviour related to the levels oftheir mental workload were measured to assess performance differences. Moreover, the changes of their mental workload were evaluated using the Modified Cooper and Harper scale (MCH-scale) and the spectral peak's power of the heart rate variability (HRV) near 0·1 beat- t region (0'1 HRV power).
2. Methods 2.1. Subjects The subjects were five experienced male pilots employed by civil airlines. Their average age was 42·2 ± 3·7 years old, and their average experience 8689 ± 2318 h of total flight time. In the experiment, the subjects used flight simulators which correspond to the type of aircraft which they usually handle. The flight simulators used were a Boeing 767 (B767) for three subjects, and a Boeing 747-300 (B747) for the remaining two. 2.2. Flight tasks The simulated flight tasks were the same as those which the pilots have to perform in their everyday commercial flights. In addition, pilots were exposed to abnormal situations (i.e., taking off with engine failure, landing with engine failure, and landing
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
722
Y. Itoh
with asymmetrical flaps) without any notice. At the same time, flight tasks in a normal situation were also performed to compare the measured values under abnormal situation with those under normal condition. The total time ofthis experiment for each subject was about 2 h. In this experiment, pilots alone were subjects, even though co-pilots and flight engineers (in the experiment using the B747 flight simulator) were involved to support the ordinary flight task. Although this study focused on the characteristics of pilots' scanning behaviour in the flight phases 'take-off' and 'landing', the subjects performed the entire process of a flight, from taking-off to landing, in order to simulate everyday commercial flights. This study defines the flight phase 'take-off' as the interval between the time when the aircraft velocity reaches the level (VR) where the airplane cannot stop taking-off and the time when the aircrew begins to go through the 'post-take-off checklist', while the 'landing phase' is defined as the interval between the time when the crew complete the 'landing checklist' and the 'time of touchdown'. If the flight phase is longer than 180s, the take-off phase is defined as the interval of 180s from VR. Analogously, ifthe landing phase is longer than 180s, it is defined as the interval of 180 s before touchdown. 2.3. Measurements 2.3.1. Eye movements: Eye movements were recorded by an eye mark recorder (EMR-V
manufactured by NAC Co. Ltd) using the ophthalmography method. In consideration of the shortest fixation duration, the definition of 'fixation' refers to the pilot's eyeball position fixed for more than, or equal to, 0·165 s (5 frames). Wierwille et al. (1985) reported that the proportion of fixation duration for the instruments might be related to the mental workload; the higher the cognitive mental workload, the longer the proportion of the fixation duration for instrument observation. In the present study, when the abnormal situation occurred, the changes of this proportion for each instrument, and the average fixation duration for all instruments, were investigated. The scanning area was classified into nine groups to examine the visual scan of instruments, viz., indicators for engines' condition; ® control panel; (J) an overhead panel; ® windows (for viewing outside); and ® others. The total fixation duration of each instrument group is required at every flight phase. Moreover, the entropy rate is introduced as a measurement of scan randomness for these instrument groups (Tole et al. 1983). D
Entropy Rate=
L [(E/Emax)/DT;]
(1)
i=1
where, D
E= -
LP
i= 1
E Pi DT; D
the the the the
ilog 2
Pi
observed average entropy; probability of occurrence of the sequence i; average fixation duration for the ith sequence in the flight phase; number of different sequences in the scan.
Because the instrument group at which the pilot views may depend upon a group which was viewed previously, it is important to record a sequence of instrument groups
Eye movement and mental workload in pilots
723
(A) Sample of Seanpath for the Instrument Groups (0= 5/24 • (2)= 7/24 • (3)= 1/24 • (4)= 1/24 •(5)= 4/24 • (6)= 2/24 • (7)= 1/24 • (8)= 0 • (9)= 3/24 (B) ProbabiJitv of the occurrence of one-length Sequences (I. 2)= 4/23 • (2.5)= 4/23 • (5.6)= 2/23 • (5.9)= 2/23 • (9.0= 2/23 •
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
(4.9)= 1/23 • (7.2)= 1/23 • (2.1)= 1/23 • (6.1)= 1/23 • (9.2)= 1/23 • (1.3)= 1/23 • (2.4)= 1/23 • (3.2)= 1/23 • (C) Probabi I i tf of the occurrence of two-length Sequences
Figure 3. The method of the calculation for the probability of the occurring sequences. Note: (n) and (n,m) denote the sequence of the instrument groups. monitored in turn, based on how a pilot watches instruments. If a pilot watches two groups of instruments in a sequence of monitoring, the length ofthe sequence is defined to be 2. This study examined the entropy rates in the sequences of the instrument groups whose length were I and 2. The number of different sequences in the scan for one-length sequence is 9, and that for two-length sequence is 72. Entropy (E) has the unit of bits/sequence. A higher entropy value represents more scan randomness. Furthermore, considering the fact that the average duration of each sequence varies, the entropy rate is normalized by this average. This index unit is bits/so If the pilot repeatedly scans only one sequence, the entropy rate would be minimum value (= 0). Conversely, the value would be maximum if -the pilot views all of the sequences with same probability, and if the average fixation duration is equal to 0 s, i.e., the higher the entropy rate, the more random the scan performed. The maximum value for an one-length sequence becomes 1/0·165 [bits/s], and that for a two-length sequence becomes 1/(0'165 x 2) [bits/s], In the present study, the entropy rate was calculated by using all scan paths performed, considering the order of sequence. Figure 3 shows the calculation method for the probability of the occurring sequences. 2.3.2. Subjective mental workload: After performing the flight, subjective ratings were given to every flight phase by Modified Cooper and Harper scale (MCH-scale) which covers the to-point decision tree. The MCH-scale is the most popular and widely accepted for pilot mental workload measurement' (Skipper 1986). The higher the ratings of the scale, the higher the subjective mental workload. Since the subjects were well-trained pilots, they were expected to be able to evaluate their mental workload appropriately. It is assumed that the subjective mental workload could be accurately measured by this rating scale. 2.3.3. Heart rate variability (H R V): In addition to the subjective ratings, HRV is used to measure an operator's mental workload. This measurement is obtained by analysis of the variability of successive R-R intervals from an electrocardiogram. Some researches showed that HRV can accurately measure the operator's mental workload. Moreover,
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
724
Y. Itoh
it was reported that the spectral peak's power of H RV near O· 1 beat - 1 region (about lOs cyclic component) systematically decreases in response to an increase in mental workload (Hyndman 1975, Egelund 1982, Aasman et al. 1987, Vincent et al. 1987, Itoh 1988). In this experiment, we employed that power value (0·1 HRV power) as an objective mental workload measurement. Throughout the experiment, the subjects' electrocardiograms were recorded with a small portable recorder (DMC-3125 Nihon Koden Kougyou Co. Ltd). The signals of R-R intervals were obtained from electrocardiograms. In order to evaluate the changes of the HRV caused by flight situations, the subjects' electrocardiograms were recorded as control while they sat in a quiet state for 3 min before the 'flight'. The power spectrum is calculated from the R-R interval signal using the 13 order auto regression model. The 0·1 HRV power was investigated then. In this paper, the unit of the spectral power is defined as follows, (2)
[dB] where, Pu the spectral peak's power near 0·1 beat
3,
-1
region of ith subject injth flight phase.
Results
3.1. The changes of subjective ratings and heart rate variability Based on the observations on subjective ratings and 0·1 HRV power, it was noted that the mental workload was higher in abnormal situation (figure 4).This was shown by the obtained ratings of M CH -scale which were estimated as significantly higher under abnormal situtations than those under normal situations (t = 5·638, P '" ~ a.15
0.10
0.10
C>
~
w
w
0.05
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
9.25
'" 0.26
r
0.06
B767A B767C B747B B767B B747A SUBJECT (A)
I
B767A B767C B747 B767B B747 SUBJECT
TAK I NO OFF
18)
lANDING
Figure 8. The entropy rates of one-length sequences for the instrument groups under normal and abnormal situation. Note: B767A means Subject A using Boeing 767 flight simulator. B767B means Subject B using Boeing 767 flight simulator. B767C means Subject C using Boeing 767 flight simulator. B747A means Subject A using Boeing 747 flight simulator. B747B means Subject B using Boeing 747 flight simulator. blt!ll/s.c
•. 6
w 8.5
w
t-
t-
"
"
•••
. ...
&
&
> 0
&
tZ
w
•. 3
•• 3
BTOTA elG'B BTGTe
BT~7A
814TB
SUBJECT IS)
LANDING
Figure 9. The entropy rates of two-length sequences for the instrument groups under normal and abnormal situations. Note: B767A means Subject A using Boeing 767 flight simulator. B767B means Subject B using Boeing 767 flight simulator. B767C means Subject C using Boeing 767 flight simulator. B747A means Subject A using Boeing 747 flight simulator. B747B means Subject B using Boeing 747 flight simulator.
instrument groups. Hence, the attitude director indicator might be an important instrument for pilot tasks. Moreover, when the abnormal situation occurred, the proportion for particular instrument groups became larger. These particular instruments for 8767 pilots were the air speed indicator in the take-off phase, and the attitude director indicator in the landing phase; for 8747 pilots, these were the attitude director indicator in the take-off phase, and the window in the landing phase. Furthermore, figurcIO shows that when the abnormal situation occurred, the changes in scanning behaviour for the 8747 pilots were larger than those of 8767 pilots. For example, the changes of proportions of the particular instruments for 8747 pilots are larger than those for 8767 pilots.
729
Eye movement and mental workload in pilots B
20
40
(A-II B
80
tBB
HIKING-OFF UU 4e
20
•B
sa
60
(~-2)
,
NORMOL~ ABNORMA 18-1)
i
~:~~
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
•
001 \01 I NDO,",
lii !I
60
•
80
10B
LAND ING
40
,....................... B
20
BlII
80
'B B74Hl
HSI
III
AIR SPEED
OTHERS
•
MDUING
2)
•
1 BO
~~
if
:
TAKING-OFF
'B>
0
49
B7B7A
ui
80
20
......
Vj
Li=lNDING
o
ALTIMETER
Figure 10. Examples of the proportion ofthe fixation duration for each instrument group to the total fixation duration within the sampling time. Note: (I) ADI: the proportion of the fixation duration for the attitude director indicator; HSI: that for the horizontal situation indicator; AIRSP: that for the air speed indicator; ALTI: that for the altimeter indicator; WINDOW: that for the window; OTHERS: that for the overhead panel, control panel, and the other instruments; MOVE: the proportion of time during the eyeball points are moving. (2) B767A means Subject A using Boeing 767 flight simulator. B747A means Subject A using Boeing 747 flight simulator.
(A) Boeing 1~1-300 Figure II.
(B) Boeing 761 Attitude director indicators.
4. Discussion The figures of pilot scanning pattern (see figures 5 and 6) and of the proportions of fixation duration for each instrument (see figure 8) show that the pilots stared at the attitude director indicator more frequently than at other instruments. Hence, the attitude director indicator might be one of the most important instruments for pilot tasks. However, figure 8 shows that in the landing phase, the 8747 pilots often stared at the window giving them an appropriate view outside. This behaviour is caused by the design of the 8747's attitude director indicator which is different from that of the 8767's indicator. Figure II represents the design of the attitude director indicators in the 8747 and 8767. The B747's attitude director indicator is an earlier electromechanical
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
730
Y. Itoh
display. With such a display, 8747-pilots often need to look outside to maintain the attitude of aircraft, hence the proportion for viewing the window becomes large. For the 8767 pilots, however, the pilot can use the flight director instead. 8ecause the flight director is displayed in the attitude director indicator, the proportion of fixation duration for the attitude director indicator ofthe 8767 pilot amounts to 80% in landing phase. Similarly, in the take-off phase, the proportion of fixation duration for the attitude director indicator of the 8767 pilots has the largest value of the proportion for other instruments. With regard to scanning behaviour, the values of the average fixation duration for all instruments of 8767 pilots were longer than those of 8747 pilots. This characteristic of8767 pilots is verified by the result of the entropy rate. The values of the entropy rates for 8767 pilots were smaller than those for the 8747's pilots. This indicates that 8767 pilots performed partial scanning, this is to say, they tended to monitor a certain instrument, namely the attitude director indicator. Therefore, it is suggested that by the introduction of the integrated displays, the pilots tend to stare at these integrated displays more frequently and that the 'gaze-type scanning' might be one of the characteristics of the pilot scanning behaviour in the glass cock pit. On the other hand, the subjective rating and 0·1 HRV power indicate that the pilot carries the high mental workload when an abnormal situation occurred. However, the differences of the mental workload between 8767 and 8747 pilots could not be observed. When abnormal situations occurred, the proportion of the total fixation duration increased for certain instruments (see figure 10). The relationship between this proportion and the subjective mental workload is shown in figure 12. It is found that the proportion value increases in response to the increase of mental workload. The pilots said that if the engine failure happens during take-off, maintaining speed and attitude is imperative. In this situation, the pilot is placed in a high mental workload condition, and it makes pilots tend to monitor the instruments concerned. The tendencies of monitoring particular instruments in landing phase could be explained by the same reasons as those in the take-off phase. Furthermore, figure 10 and II show that in the abnormal situation, the changes of the scanning behaviour for 8767 pilots were smaller than those of8747 pilots. It may be
..•
.• 4 ~ 3 :;;
I NORMAL NORMAL ABNORMAL ABNORMAL TAK I NO OFF LAND I NO
NORMAL NORMAL ABNORMAL AIINORnAl TAKING OFF LANDING
PHASE (A)
818TA
PHASE ISl
B1HA
Figure 12. The relationship between the proportion of fixation duration for the particular instrument to the total fixation duration and subjective mental workload. Note: B767A meansSubject A usingBoeing 767 flight simulator. B747A meansSubject A usingBoeing 747 flight simulator.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
Eye movement and mental workload in pilots
731
implied that the application of the integrated displays helps 8767 pilots to be able to obtain sufficient information more easily than 8747 pilots, even under abnormal situations. From the result as above, it is suggested that there is no difference in mental workload between pilots using integrated CRT displays and pilots using earlier, electromechanical displays. However, in terms of the ease of obtaining information from instruments, the integrated displays using the CRT might be more preferable for pilot tasks. Given that the subjects in our experiment were experienced pilots, accustomed to flying aircraft using electromechanical display, it is perhaps unsurprising that no differences in mental workload between 8767 pilots and 8747 pilots were observed. However, if the subjects in this experiment had an experience of flight operations only in the glass cockpit, their mental workload could be higher than the workload in the electromechanical display environment, because the introduction of the automatic systems and the simplification of the task might cause a decrease in the operators' precaution for emergencies, or a decrease of the operators' technical capacity when manual control is needed. Unfortunately, we cannot draw any conclusions from the results of our experiment. However, it is predicted that this aspect of an increase in mental workload under an abnormal situation will become a matter of immediate concern in the future. On the other hand, the variation of the visual environment in the cockpit between the air routes, such as illuminance and brightness, is large. That is to say, when the aircraft is flying to the direction of the sun, illuminance is very high. Conversely, throughout the flight in the dark, illuminance is very low. It is assumed that these conditions could affect pilot performance while using CRT display. Furthermore, the work hours and work content of pilots are also significantly different from those of office workers, even though both pilots and office workers use CRT displays. Regulations concerning the environment for office VDT work cannot be applied to flying. In the near future, it will be necessary to discuss the problem of the pilot fatigue caused by the introduction of the CRT displays in the cockpit.
Acknowledgements This work was supported by the pilots who took part in our experiments, and thanks are due to them. Furthermore, the authors acknowledge Mr H. Asai, Mr Y. Yasumoto, Mr R. Komatsu, Mr T. Suzuki and Aeromedical Research Center staff for their co-operation. References AASMAN, J., MULDER, G. and MULDER, L. 1. M. 1987, Operator effort and the measurement of heart rate variability, Human Factors, 29,161-170. EGELUND, N. 1982, Heart rate variability as an indicator of driver fatigue, Ergonomics, 25, 663-672.
CARAUX, D. and WANNER, J. C. 1979, Pilot workload in the aircraft of the future, NASA Conference Publication 2103. HAWKINS, F. H. (ed.), 1987, Human Factors in Flight (Gower Technical Press, Aldershot). HAYASHI, Y. (ed.), 1988, The Effect oj Lighting in a Cockpit on the Pilot's Sight Function (Japan Aero Medical Research Center, Tokyo); (in Japanese). HYNDMAN, B. and GREGORY, 1. R. 1975, Spectral analysis of sinus arrhythmia during mental loading, Ergonomics 18, 225-270.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
732
Y. ltoh
ITOH, Y. 1988, The relation between the changes of biophysiological reactions and subjective mental workload in typing tasks under time .pressure, Japan Ergonomics, 24, 253-260; (in Japanese). SAYERS, B. M. 1971, Analysis of heart rate variability, Ergonomics, 16,117-132. SKIPPER, J. H., RIEGER, C. A. and WIERWILLE, W. W. 1986, Evaluation of decision-tree rating scales for mental workload estimation, Ergonomics, 29, 585-599. SPYKER, D. A., STACKHOUSE, S. P., KHALAFALLA, A. S. and McLANE, R. C. 1971, Development of techniques for measuring pilot workload, NASA Contractor Report 1888. TOLE, J. R., STEPHENS, A. T., VIVAUDOU, M., EpHRA TH, A. and YOUNG, L. R. 1983,Visual scanning behaviour and pilot workload, NASA Contractor Report 3717. VINCENT, K. J., THORNTON, D. C. and MORAY, N.1987, Spectral analysis of sinus arrhythmia: a measure of mental effort, Human Factors, 29, 171-182. WIERWILLE, W. W., RAHIMI, M. and CASALI, J. G. 1985, Evaluation of 16 measures of mental workload using a simulated flight task emphasizing mediational activity, Human Factors, 27, 489-502. Manuscript received 14 June 1989. Manuscript accepted 27 November 1989.
Cet article presente une experience destinee a etudier les caracteristiques de scrutage des pilotes sur un ecran de visualisation, ainsi que les modifications survenant dans ces caracteristiques lorsque Ie pilote doit affronter une situation anormale. Un groupe de cinq pilotes entraines dcvait effcctuer des taches de pilotage sous deux modalites differentes, dans des simulateurs de vol, l'une en situation normale, I'autre en situation anormale. On simulait Ie vol soit d'un Boeing 767-300 (B747) utilisant un dispositif electromecanique, soit d'un Boeing 767 (B767) utilisant un dispositif ecran integre, Les resultats montrent que les pilotes sur B767 tendaient fixer Ie regard sur l'indicateur de position present dans Ie dispositif ecran. 11 est done possible que ce type de strategic d'observation visuelle caracterise Ie vol avec un dispositif integre. Cepedant il n'y a aucune difference dans la charge mentale entre les pilotes de B747 et ceux de B767, a en juger d'apres Ics evaluations subjectives et la variabilite de frequence cardiaque. Toutefois, lorsqu'une situation anormale se presente, il apparait que les modifications dans la strategic de scrutage sont moindres chez les pilotes de B767 que chez les pilotes de B747. On conclut que l'utilisation du dispositif integre permet au pilote d'obtenir plus facilement l'information souhaite et ceci meme dans les situations anormales.
a
Dieser Beitrag beschreibt ein Experiment, das die Eigenschaften des Abtastverhaltens von Piloten untersucht, wiihrend sie integrierte CRT-Displays benutzen, sowie die Anderung der Eigenschaften, wenn die Piloten einer anomalen Situation gegeniiberstehen. Ais Versuchspersonen wurden fiinf erfahrene Piloten eingesetzt. Sie fiihrten zwei Arten von Flugaufgaben unter normalen und unter anomalen Bedingungen in Flugsimulatoren mit Standardausstattungen aus. Die Flugsimulatoren waren zum einen fiir eine Boeing 747-300 (B747), die elektro-mechanische Displays benutzten, und zum anderen fiir eine Boeing 767 (B767), die mit integrierten CRT-Displays ausgestattet waren. Die Ergebnisse zeigten, daB die B767-Piloten dazu neigten, die 'attitude-director' Anzeige zu fixieren, die in dem intergrierten CRT-Display angezeigt wurde. Es wurde angenommen, daB das Abtastverhalten mit betontem Fixieren eine den Eigenschaften des Abtastverhaltens von Piloten in Cockpits sein konnte, die das integrierte CRT-Display benutzen. Unter Verwendung von subjektiven Beurteilungen und der Herzschlagvariabilitiit zur Messung der mentalen Beanspruchung wurden keine Unterschiede der mentalen Beanspruchung zwischen den B767 und den B747 Piloten beobachtet. Es wurde jedoch herausgefunden, daB in anomalen Situationen die Anderung der Abtastmuster fiir B767 kleiner als die der B747 Piloten waren. Es wird geschlossen, daB der Einsatz von integrierten Displays den Piloten hilft, ausreiehend Informationen leichter zu erhalten als mit elektro-mechanischen Displays. Dies gilt aueh fiir anomale Situationen.
Ergonomics Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/terg20
The ergonomic evaluation of eye movement and mental workload in aircraft pilots a
a
b
Yasuko Itoh , Yoshio Hayashi , Ippei Tsukui & Susumu Saito
c
a
Department of Administration Engineering , Keio University , 3-14-1, Hiyoshi, Kohoku-ku, Yokohama, 223, Japan. b
Japan Aeromedical Research Center , 2-3-1, Haneda Airport, Ohta-ku, Tokyo, 144, Japan.
c
National Institute of Industrial Health , Nagao 6, Tama-ku, Kawasaki, 214, Japan. Published online: 27 Mar 2007.
To cite this article: Yasuko Itoh , Yoshio Hayashi , Ippei Tsukui & Susumu Saito (1990) The ergonomic evaluation of eye movement and mental workload in aircraft pilots, Ergonomics, 33:6, 719-732, DOI: 10.1080/00140139008927181 To link to this article: http://dx.doi.org/10.1080/00140139008927181
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
ERGONOMICS, 1990, VOL. 33, No.6, 719-733
The ergonomic evaluation of eye movement and mental workload in aircraft pilots YASUKO ITOH and YOSHIO HAYASHI Department of Administration Engineering, Keio University, 3-14-1, Hiyoshi, Kohoku-ku, Yokohama 223, Japan
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
IpPEI TSUKUI Japan Aeromedical Research Center, 2-3-1, Haneda Airport, Ohta-ku, Tokyo 144, Japan SUSUMU SAITO National Institute of Industrial Health, Nagao 6, Tama-ku, Kawasaki 214, Japan Keywords: Eye movement;Visual scanning; Entropy; Integrated display; Heart rate
variability; Mental workload. This paper presents an experiment which examines characteristics of pilots' scanning behaviour when using integrated CRT displays, and the changes in characteristics when pilots face abnormal situations. The subjects were five experienced pilots. They performed two modes of flight tasks, under normal and abnormal situations, in flight simulators with standard settings. The flight simulators were for a Boeing 747-300(B747), which made use of electromechanical displays, and for a Boeing 767 (B767), equipped with integrated CRT displays. The results showed that the B767pilots tended to gaze at the attitude director indicator which was displayed in the integrated CRT display. It was assumed that 'gaze-type scanning' might be one of the characteristics of pilots' scanning behaviour in cockpits which use the integrated display. By employing subjective ratings and heart rate variability to measure mental workload, no differences in mental workload between the B767pilots and the B747pilots were observed. However, in abnormal situations, the changes in scanning pattern for B767pilots were found to be smaller than those of the B747 pilots. It is concluded that the application of integrated displays helps pilots to obtain sufficient information more easily than electromechanical displays do, even under abnormal situations.
1. Introduction The increasing complexity of navigation procedures and more stringent safety standards for civil transport aircraft have led to a proliferation of instruments and information systems to enable pilots to obtain the essential information to perform flying operations safely and satisfactorily. On the other hand, statistics reveal that the factor which most contributes to total loss accidents caused by crew behaviour is inadequate judgement and misoperation (see figure I). Therefore, information display in cockpits to enable the pilot to make rapid and correct decisions is imperative. Some studies focus on the instrument which provides an integrated presentation of the flight situation and which is adaptable to the different phases of flight. In the early 1980s, an introduction ofcathode-ray tube (CRT) flight displays in civil aircraft marked a watershed in the evolution of information displays in cockpits. The use of CRT 0014-0139/90 $3-00 © 1990 Taylor & Francis Ltd.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
720
Figure I.
Y. Itoh
Factors influencing total loss accidents caused by the crew for the passenger jet planes, worldwide (1959-1983); Source: ICAO Bulletin.
(~)
~oeing
m
.J •..• J. _t.".t_
iu::":U:
(C)
~oeing 141-4~U
Figure 2. Examples of cockpit instrument panels.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
Eye movement and mental workload in pilots
721
displays achieved an integration of instruments, so facilitating the more effective utilization of high priority panel space and greater flexibility. Furthermore, timesharing of the high priority panel space is thereby realized, and information which is not relevant in any particular phase of flight is removed from view altogether. As an example of the integrated displays and electromechanical display, figure 2 shows the control panels of the Boeing 747-300 (8747), the Boeing 767 (8767), and the Boeing 747-400(8747-400). While the earlier electromechanical instruments are used in 8747 cockpits, the new technology of integrated instrument systems using CRT displays has been introduced into 8767 and 8747-400 cockpits. The displays in the 8747-400 cockpit are more advanced than those in the 8767 cockpit, in terms of integration of instruments. For example, the Primary Flight Display in the 8747-400 cockpit integrates the attitude director indicator and the horizontal situation indicator with other instruments into a single display. The 8767 cockpit itself is considerably more integrated than the 8747 cockpit. In order to compare the B767 with the B747, one may take the attitude director indicator as an example of integrated devices. Although the Electro Attitude Director Indicator (EADI) in the B767 cockpit looks somewhat similar to the electromechanical display of the altitude director indicator in the B747 cockpit, the EADI has a more advanced function than the latter, because the EADI can display a tracking target which the plane must head towards. Thus, the EADI provides enhanced support to pilots. Both the B767 and the B747 flight simulators were used in this experiment. It is claimed that the operational benefit of integrated displays is a reduction in the scanning area for essential information, so that the pilots can easily understand flight conditions. On the other hand, pilot performance and workload might be changed by the introduction of such integrated displays. In particular, pilots' eye movements might be influenced. Thus, it is important to ascertain whether or not the introduction of the integrated displays is preferable for pilots. This paper examines the differences in scanning behaviour under conditions of high mental workload between pilots using an integrated display, and pilots using the earlier electromechanical display. The experiment was conducted in B767 and 8747 flight simulators, with standard settings. Pilots were exposed to abnormal situations without any notice, and the changes of scanning behaviour related to the levels oftheir mental workload were measured to assess performance differences. Moreover, the changes of their mental workload were evaluated using the Modified Cooper and Harper scale (MCH-scale) and the spectral peak's power of the heart rate variability (HRV) near 0·1 beat- t region (0'1 HRV power).
2. Methods 2.1. Subjects The subjects were five experienced male pilots employed by civil airlines. Their average age was 42·2 ± 3·7 years old, and their average experience 8689 ± 2318 h of total flight time. In the experiment, the subjects used flight simulators which correspond to the type of aircraft which they usually handle. The flight simulators used were a Boeing 767 (B767) for three subjects, and a Boeing 747-300 (B747) for the remaining two. 2.2. Flight tasks The simulated flight tasks were the same as those which the pilots have to perform in their everyday commercial flights. In addition, pilots were exposed to abnormal situations (i.e., taking off with engine failure, landing with engine failure, and landing
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
722
Y. Itoh
with asymmetrical flaps) without any notice. At the same time, flight tasks in a normal situation were also performed to compare the measured values under abnormal situation with those under normal condition. The total time ofthis experiment for each subject was about 2 h. In this experiment, pilots alone were subjects, even though co-pilots and flight engineers (in the experiment using the B747 flight simulator) were involved to support the ordinary flight task. Although this study focused on the characteristics of pilots' scanning behaviour in the flight phases 'take-off' and 'landing', the subjects performed the entire process of a flight, from taking-off to landing, in order to simulate everyday commercial flights. This study defines the flight phase 'take-off' as the interval between the time when the aircraft velocity reaches the level (VR) where the airplane cannot stop taking-off and the time when the aircrew begins to go through the 'post-take-off checklist', while the 'landing phase' is defined as the interval between the time when the crew complete the 'landing checklist' and the 'time of touchdown'. If the flight phase is longer than 180s, the take-off phase is defined as the interval of 180s from VR. Analogously, ifthe landing phase is longer than 180s, it is defined as the interval of 180 s before touchdown. 2.3. Measurements 2.3.1. Eye movements: Eye movements were recorded by an eye mark recorder (EMR-V
manufactured by NAC Co. Ltd) using the ophthalmography method. In consideration of the shortest fixation duration, the definition of 'fixation' refers to the pilot's eyeball position fixed for more than, or equal to, 0·165 s (5 frames). Wierwille et al. (1985) reported that the proportion of fixation duration for the instruments might be related to the mental workload; the higher the cognitive mental workload, the longer the proportion of the fixation duration for instrument observation. In the present study, when the abnormal situation occurred, the changes of this proportion for each instrument, and the average fixation duration for all instruments, were investigated. The scanning area was classified into nine groups to examine the visual scan of instruments, viz., indicators for engines' condition; ® control panel; (J) an overhead panel; ® windows (for viewing outside); and ® others. The total fixation duration of each instrument group is required at every flight phase. Moreover, the entropy rate is introduced as a measurement of scan randomness for these instrument groups (Tole et al. 1983). D
Entropy Rate=
L [(E/Emax)/DT;]
(1)
i=1
where, D
E= -
LP
i= 1
E Pi DT; D
the the the the
ilog 2
Pi
observed average entropy; probability of occurrence of the sequence i; average fixation duration for the ith sequence in the flight phase; number of different sequences in the scan.
Because the instrument group at which the pilot views may depend upon a group which was viewed previously, it is important to record a sequence of instrument groups
Eye movement and mental workload in pilots
723
(A) Sample of Seanpath for the Instrument Groups (0= 5/24 • (2)= 7/24 • (3)= 1/24 • (4)= 1/24 •(5)= 4/24 • (6)= 2/24 • (7)= 1/24 • (8)= 0 • (9)= 3/24 (B) ProbabiJitv of the occurrence of one-length Sequences (I. 2)= 4/23 • (2.5)= 4/23 • (5.6)= 2/23 • (5.9)= 2/23 • (9.0= 2/23 •
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
(4.9)= 1/23 • (7.2)= 1/23 • (2.1)= 1/23 • (6.1)= 1/23 • (9.2)= 1/23 • (1.3)= 1/23 • (2.4)= 1/23 • (3.2)= 1/23 • (C) Probabi I i tf of the occurrence of two-length Sequences
Figure 3. The method of the calculation for the probability of the occurring sequences. Note: (n) and (n,m) denote the sequence of the instrument groups. monitored in turn, based on how a pilot watches instruments. If a pilot watches two groups of instruments in a sequence of monitoring, the length ofthe sequence is defined to be 2. This study examined the entropy rates in the sequences of the instrument groups whose length were I and 2. The number of different sequences in the scan for one-length sequence is 9, and that for two-length sequence is 72. Entropy (E) has the unit of bits/sequence. A higher entropy value represents more scan randomness. Furthermore, considering the fact that the average duration of each sequence varies, the entropy rate is normalized by this average. This index unit is bits/so If the pilot repeatedly scans only one sequence, the entropy rate would be minimum value (= 0). Conversely, the value would be maximum if -the pilot views all of the sequences with same probability, and if the average fixation duration is equal to 0 s, i.e., the higher the entropy rate, the more random the scan performed. The maximum value for an one-length sequence becomes 1/0·165 [bits/s], and that for a two-length sequence becomes 1/(0'165 x 2) [bits/s], In the present study, the entropy rate was calculated by using all scan paths performed, considering the order of sequence. Figure 3 shows the calculation method for the probability of the occurring sequences. 2.3.2. Subjective mental workload: After performing the flight, subjective ratings were given to every flight phase by Modified Cooper and Harper scale (MCH-scale) which covers the to-point decision tree. The MCH-scale is the most popular and widely accepted for pilot mental workload measurement' (Skipper 1986). The higher the ratings of the scale, the higher the subjective mental workload. Since the subjects were well-trained pilots, they were expected to be able to evaluate their mental workload appropriately. It is assumed that the subjective mental workload could be accurately measured by this rating scale. 2.3.3. Heart rate variability (H R V): In addition to the subjective ratings, HRV is used to measure an operator's mental workload. This measurement is obtained by analysis of the variability of successive R-R intervals from an electrocardiogram. Some researches showed that HRV can accurately measure the operator's mental workload. Moreover,
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
724
Y. Itoh
it was reported that the spectral peak's power of H RV near O· 1 beat - 1 region (about lOs cyclic component) systematically decreases in response to an increase in mental workload (Hyndman 1975, Egelund 1982, Aasman et al. 1987, Vincent et al. 1987, Itoh 1988). In this experiment, we employed that power value (0·1 HRV power) as an objective mental workload measurement. Throughout the experiment, the subjects' electrocardiograms were recorded with a small portable recorder (DMC-3125 Nihon Koden Kougyou Co. Ltd). The signals of R-R intervals were obtained from electrocardiograms. In order to evaluate the changes of the HRV caused by flight situations, the subjects' electrocardiograms were recorded as control while they sat in a quiet state for 3 min before the 'flight'. The power spectrum is calculated from the R-R interval signal using the 13 order auto regression model. The 0·1 HRV power was investigated then. In this paper, the unit of the spectral power is defined as follows, (2)
[dB] where, Pu the spectral peak's power near 0·1 beat
3,
-1
region of ith subject injth flight phase.
Results
3.1. The changes of subjective ratings and heart rate variability Based on the observations on subjective ratings and 0·1 HRV power, it was noted that the mental workload was higher in abnormal situation (figure 4).This was shown by the obtained ratings of M CH -scale which were estimated as significantly higher under abnormal situtations than those under normal situations (t = 5·638, P '" ~ a.15
0.10
0.10
C>
~
w
w
0.05
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
9.25
'" 0.26
r
0.06
B767A B767C B747B B767B B747A SUBJECT (A)
I
B767A B767C B747 B767B B747 SUBJECT
TAK I NO OFF
18)
lANDING
Figure 8. The entropy rates of one-length sequences for the instrument groups under normal and abnormal situation. Note: B767A means Subject A using Boeing 767 flight simulator. B767B means Subject B using Boeing 767 flight simulator. B767C means Subject C using Boeing 767 flight simulator. B747A means Subject A using Boeing 747 flight simulator. B747B means Subject B using Boeing 747 flight simulator. blt!ll/s.c
•. 6
w 8.5
w
t-
t-
"
"
•••
. ...
&
&
> 0
&
tZ
w
•. 3
•• 3
BTOTA elG'B BTGTe
BT~7A
814TB
SUBJECT IS)
LANDING
Figure 9. The entropy rates of two-length sequences for the instrument groups under normal and abnormal situations. Note: B767A means Subject A using Boeing 767 flight simulator. B767B means Subject B using Boeing 767 flight simulator. B767C means Subject C using Boeing 767 flight simulator. B747A means Subject A using Boeing 747 flight simulator. B747B means Subject B using Boeing 747 flight simulator.
instrument groups. Hence, the attitude director indicator might be an important instrument for pilot tasks. Moreover, when the abnormal situation occurred, the proportion for particular instrument groups became larger. These particular instruments for 8767 pilots were the air speed indicator in the take-off phase, and the attitude director indicator in the landing phase; for 8747 pilots, these were the attitude director indicator in the take-off phase, and the window in the landing phase. Furthermore, figurcIO shows that when the abnormal situation occurred, the changes in scanning behaviour for the 8747 pilots were larger than those of 8767 pilots. For example, the changes of proportions of the particular instruments for 8747 pilots are larger than those for 8767 pilots.
729
Eye movement and mental workload in pilots B
20
40
(A-II B
80
tBB
HIKING-OFF UU 4e
20
•B
sa
60
(~-2)
,
NORMOL~ ABNORMA 18-1)
i
~:~~
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
•
001 \01 I NDO,",
lii !I
60
•
80
10B
LAND ING
40
,....................... B
20
BlII
80
'B B74Hl
HSI
III
AIR SPEED
OTHERS
•
MDUING
2)
•
1 BO
~~
if
:
TAKING-OFF
'B>
0
49
B7B7A
ui
80
20
......
Vj
Li=lNDING
o
ALTIMETER
Figure 10. Examples of the proportion ofthe fixation duration for each instrument group to the total fixation duration within the sampling time. Note: (I) ADI: the proportion of the fixation duration for the attitude director indicator; HSI: that for the horizontal situation indicator; AIRSP: that for the air speed indicator; ALTI: that for the altimeter indicator; WINDOW: that for the window; OTHERS: that for the overhead panel, control panel, and the other instruments; MOVE: the proportion of time during the eyeball points are moving. (2) B767A means Subject A using Boeing 767 flight simulator. B747A means Subject A using Boeing 747 flight simulator.
(A) Boeing 1~1-300 Figure II.
(B) Boeing 761 Attitude director indicators.
4. Discussion The figures of pilot scanning pattern (see figures 5 and 6) and of the proportions of fixation duration for each instrument (see figure 8) show that the pilots stared at the attitude director indicator more frequently than at other instruments. Hence, the attitude director indicator might be one of the most important instruments for pilot tasks. However, figure 8 shows that in the landing phase, the 8747 pilots often stared at the window giving them an appropriate view outside. This behaviour is caused by the design of the 8747's attitude director indicator which is different from that of the 8767's indicator. Figure II represents the design of the attitude director indicators in the 8747 and 8767. The B747's attitude director indicator is an earlier electromechanical
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
730
Y. Itoh
display. With such a display, 8747-pilots often need to look outside to maintain the attitude of aircraft, hence the proportion for viewing the window becomes large. For the 8767 pilots, however, the pilot can use the flight director instead. 8ecause the flight director is displayed in the attitude director indicator, the proportion of fixation duration for the attitude director indicator ofthe 8767 pilot amounts to 80% in landing phase. Similarly, in the take-off phase, the proportion of fixation duration for the attitude director indicator of the 8767 pilots has the largest value of the proportion for other instruments. With regard to scanning behaviour, the values of the average fixation duration for all instruments of 8767 pilots were longer than those of 8747 pilots. This characteristic of8767 pilots is verified by the result of the entropy rate. The values of the entropy rates for 8767 pilots were smaller than those for the 8747's pilots. This indicates that 8767 pilots performed partial scanning, this is to say, they tended to monitor a certain instrument, namely the attitude director indicator. Therefore, it is suggested that by the introduction of the integrated displays, the pilots tend to stare at these integrated displays more frequently and that the 'gaze-type scanning' might be one of the characteristics of the pilot scanning behaviour in the glass cock pit. On the other hand, the subjective rating and 0·1 HRV power indicate that the pilot carries the high mental workload when an abnormal situation occurred. However, the differences of the mental workload between 8767 and 8747 pilots could not be observed. When abnormal situations occurred, the proportion of the total fixation duration increased for certain instruments (see figure 10). The relationship between this proportion and the subjective mental workload is shown in figure 12. It is found that the proportion value increases in response to the increase of mental workload. The pilots said that if the engine failure happens during take-off, maintaining speed and attitude is imperative. In this situation, the pilot is placed in a high mental workload condition, and it makes pilots tend to monitor the instruments concerned. The tendencies of monitoring particular instruments in landing phase could be explained by the same reasons as those in the take-off phase. Furthermore, figure 10 and II show that in the abnormal situation, the changes of the scanning behaviour for 8767 pilots were smaller than those of8747 pilots. It may be
..•
.• 4 ~ 3 :;;
I NORMAL NORMAL ABNORMAL ABNORMAL TAK I NO OFF LAND I NO
NORMAL NORMAL ABNORMAL AIINORnAl TAKING OFF LANDING
PHASE (A)
818TA
PHASE ISl
B1HA
Figure 12. The relationship between the proportion of fixation duration for the particular instrument to the total fixation duration and subjective mental workload. Note: B767A meansSubject A usingBoeing 767 flight simulator. B747A meansSubject A usingBoeing 747 flight simulator.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
Eye movement and mental workload in pilots
731
implied that the application of the integrated displays helps 8767 pilots to be able to obtain sufficient information more easily than 8747 pilots, even under abnormal situations. From the result as above, it is suggested that there is no difference in mental workload between pilots using integrated CRT displays and pilots using earlier, electromechanical displays. However, in terms of the ease of obtaining information from instruments, the integrated displays using the CRT might be more preferable for pilot tasks. Given that the subjects in our experiment were experienced pilots, accustomed to flying aircraft using electromechanical display, it is perhaps unsurprising that no differences in mental workload between 8767 pilots and 8747 pilots were observed. However, if the subjects in this experiment had an experience of flight operations only in the glass cockpit, their mental workload could be higher than the workload in the electromechanical display environment, because the introduction of the automatic systems and the simplification of the task might cause a decrease in the operators' precaution for emergencies, or a decrease of the operators' technical capacity when manual control is needed. Unfortunately, we cannot draw any conclusions from the results of our experiment. However, it is predicted that this aspect of an increase in mental workload under an abnormal situation will become a matter of immediate concern in the future. On the other hand, the variation of the visual environment in the cockpit between the air routes, such as illuminance and brightness, is large. That is to say, when the aircraft is flying to the direction of the sun, illuminance is very high. Conversely, throughout the flight in the dark, illuminance is very low. It is assumed that these conditions could affect pilot performance while using CRT display. Furthermore, the work hours and work content of pilots are also significantly different from those of office workers, even though both pilots and office workers use CRT displays. Regulations concerning the environment for office VDT work cannot be applied to flying. In the near future, it will be necessary to discuss the problem of the pilot fatigue caused by the introduction of the CRT displays in the cockpit.
Acknowledgements This work was supported by the pilots who took part in our experiments, and thanks are due to them. Furthermore, the authors acknowledge Mr H. Asai, Mr Y. Yasumoto, Mr R. Komatsu, Mr T. Suzuki and Aeromedical Research Center staff for their co-operation. References AASMAN, J., MULDER, G. and MULDER, L. 1. M. 1987, Operator effort and the measurement of heart rate variability, Human Factors, 29,161-170. EGELUND, N. 1982, Heart rate variability as an indicator of driver fatigue, Ergonomics, 25, 663-672.
CARAUX, D. and WANNER, J. C. 1979, Pilot workload in the aircraft of the future, NASA Conference Publication 2103. HAWKINS, F. H. (ed.), 1987, Human Factors in Flight (Gower Technical Press, Aldershot). HAYASHI, Y. (ed.), 1988, The Effect oj Lighting in a Cockpit on the Pilot's Sight Function (Japan Aero Medical Research Center, Tokyo); (in Japanese). HYNDMAN, B. and GREGORY, 1. R. 1975, Spectral analysis of sinus arrhythmia during mental loading, Ergonomics 18, 225-270.
Downloaded by [Nanyang Technological University] at 01:02 28 April 2015
732
Y. ltoh
ITOH, Y. 1988, The relation between the changes of biophysiological reactions and subjective mental workload in typing tasks under time .pressure, Japan Ergonomics, 24, 253-260; (in Japanese). SAYERS, B. M. 1971, Analysis of heart rate variability, Ergonomics, 16,117-132. SKIPPER, J. H., RIEGER, C. A. and WIERWILLE, W. W. 1986, Evaluation of decision-tree rating scales for mental workload estimation, Ergonomics, 29, 585-599. SPYKER, D. A., STACKHOUSE, S. P., KHALAFALLA, A. S. and McLANE, R. C. 1971, Development of techniques for measuring pilot workload, NASA Contractor Report 1888. TOLE, J. R., STEPHENS, A. T., VIVAUDOU, M., EpHRA TH, A. and YOUNG, L. R. 1983,Visual scanning behaviour and pilot workload, NASA Contractor Report 3717. VINCENT, K. J., THORNTON, D. C. and MORAY, N.1987, Spectral analysis of sinus arrhythmia: a measure of mental effort, Human Factors, 29, 171-182. WIERWILLE, W. W., RAHIMI, M. and CASALI, J. G. 1985, Evaluation of 16 measures of mental workload using a simulated flight task emphasizing mediational activity, Human Factors, 27, 489-502. Manuscript received 14 June 1989. Manuscript accepted 27 November 1989.
Cet article presente une experience destinee a etudier les caracteristiques de scrutage des pilotes sur un ecran de visualisation, ainsi que les modifications survenant dans ces caracteristiques lorsque Ie pilote doit affronter une situation anormale. Un groupe de cinq pilotes entraines dcvait effcctuer des taches de pilotage sous deux modalites differentes, dans des simulateurs de vol, l'une en situation normale, I'autre en situation anormale. On simulait Ie vol soit d'un Boeing 767-300 (B747) utilisant un dispositif electromecanique, soit d'un Boeing 767 (B767) utilisant un dispositif ecran integre, Les resultats montrent que les pilotes sur B767 tendaient fixer Ie regard sur l'indicateur de position present dans Ie dispositif ecran. 11 est done possible que ce type de strategic d'observation visuelle caracterise Ie vol avec un dispositif integre. Cepedant il n'y a aucune difference dans la charge mentale entre les pilotes de B747 et ceux de B767, a en juger d'apres Ics evaluations subjectives et la variabilite de frequence cardiaque. Toutefois, lorsqu'une situation anormale se presente, il apparait que les modifications dans la strategic de scrutage sont moindres chez les pilotes de B767 que chez les pilotes de B747. On conclut que l'utilisation du dispositif integre permet au pilote d'obtenir plus facilement l'information souhaite et ceci meme dans les situations anormales.
a
Dieser Beitrag beschreibt ein Experiment, das die Eigenschaften des Abtastverhaltens von Piloten untersucht, wiihrend sie integrierte CRT-Displays benutzen, sowie die Anderung der Eigenschaften, wenn die Piloten einer anomalen Situation gegeniiberstehen. Ais Versuchspersonen wurden fiinf erfahrene Piloten eingesetzt. Sie fiihrten zwei Arten von Flugaufgaben unter normalen und unter anomalen Bedingungen in Flugsimulatoren mit Standardausstattungen aus. Die Flugsimulatoren waren zum einen fiir eine Boeing 747-300 (B747), die elektro-mechanische Displays benutzten, und zum anderen fiir eine Boeing 767 (B767), die mit integrierten CRT-Displays ausgestattet waren. Die Ergebnisse zeigten, daB die B767-Piloten dazu neigten, die 'attitude-director' Anzeige zu fixieren, die in dem intergrierten CRT-Display angezeigt wurde. Es wurde angenommen, daB das Abtastverhalten mit betontem Fixieren eine den Eigenschaften des Abtastverhaltens von Piloten in Cockpits sein konnte, die das integrierte CRT-Display benutzen. Unter Verwendung von subjektiven Beurteilungen und der Herzschlagvariabilitiit zur Messung der mentalen Beanspruchung wurden keine Unterschiede der mentalen Beanspruchung zwischen den B767 und den B747 Piloten beobachtet. Es wurde jedoch herausgefunden, daB in anomalen Situationen die Anderung der Abtastmuster fiir B767 kleiner als die der B747 Piloten waren. Es wird geschlossen, daB der Einsatz von integrierten Displays den Piloten hilft, ausreiehend Informationen leichter zu erhalten als mit elektro-mechanischen Displays. Dies gilt aueh fiir anomale Situationen.
