HUMAN FACTORS, 1994,36(2),327-338Fatigue in Operational Se.docx

HUMAN FACTORS, 1994,36(2),327-338
Fatigue in Operational Settings: Examples
from the Aviation Environment
MARK R. ROSEKIND, NASA Ames Research Center,
PHILIPPA H. GANDER, San Jose State
University Foundation, DONNA L. MILLER and KEVIN B.
GREGORY, Sterling Software,
ROY M. SMITH, KERI J. WELDON, ELIZABETH L. CO, and
KAREN L. McNALLY, San
Jose State University Foundation, and J. VICTOR LEBACQZ,
NASA Ames Research Center,
Moffett Field, California
The need for 24-h operations creates nonstandard and altered
work schedules that
can lead to cumulative sleep loss and circadian disruption.
These factors can lead
to fatigue and sleepiness and affect performance and
productivity on the job. The
approach, research, and results of the NASA Ames Fatigue
Countermeasures Pro-
gram are described to illustrate one attempt to address these
issues in the aviation
environment. The scientific and operational relevance of these
factors is discussed,
and provocative issues for future research are presented.
INTRODUCTION
Today, 24-h operations are a critical com-

ponent of maintaining our technological so-
ciety. Many different types of occupations
and industries rely on round-the-clock opera-
tions, including health care, public safety,
service and manufacturing industries, mili-
tary operations. and transportation. Esti-
mates are that one in five American workers
is a shiftworker, working some form of non-
standard or altered work schedule (Office of
Technology Assessment, 1991). These 20 mil-
lion American shiftworkers are exposed to
major disruptions in their physiology, social
activities, and family lives. The principal
physiological disruption occurs in two areas:
sleep and circadian rhythms.
1 Requests for reprints should be sent to Mark R. Rose-
kind, NASA Ames Research Center. Mail Stop 262-4. Mof-
fett Field, CA 94035-1000.
Sleep is a vital physiological function, and
obtaining even 1 h less than required can af-
fect waking levels of sleepiness (Carskadon
and Dement, 1982). Sleep loss may be acute
or, if occurring continuously over time, may
result in a cumulative sleep debt (Roth,
Roehrs, Carskadon, and Dement, 1989).
The suprachiasmatic nucleus in the hypo-
thalamus is a pacemaker for 24-h physiolog-
ical and behavioral rhythms. Circadian
(about 24 h) rhythms govem sleep/wakeful-
ness, motor activity, hormonal processes,
body temperature, performance, and many
other factors. Core body temperature is often
used as a biological marker of circadian po-

sition and is related to the fluctuations seen
in sleep/wakefulness. performance, hormone
secretion, digestion, and other physiological
activities. The minimum ofthe body temper-
ature rhythm (which typically occurs at 3:00
to 5:00 a.m. daily) is associated with sleep,
328-June 1994
low motor activity, decreased performance,
and worsened mood. Disturbances of 24-h bi-
ological rhythms may be acute or continue
over long periods, resulting in chronic desyn-
chronization between different physiological
systems. (For background in this area, see
Dinges. 1989; Mistlberger and Rusak, 1989;
Monk,1989.)
Cumulative sleep loss and circadian dis-
ruption can lead to decreased waking alert-
ness, impaired performance. and worsened
mood (Bonnett, 1985; Broughton and Ogilvie,
1992). Individuals often use the word "fa-
tigue" to characterize these experiences. Es-
timates suggest that 75% of night workers ex-
perience sleepiness on every night shift, and
for 20% of them, sleepiness is so severe that
they actually fall asleep (Akerstedt, 1991). Al-
though many factors may affect the subjec-
tive report of fatigue (e.g., workload, stress,
environmental factors), the most substantial
empirical data suggest that the two principal
physiological sources of fatigue are sleep loss
and circadian disruption.

Fatigue is a concern in many operational
settings that require 24-h activities (Office of
Technology Assessment, 1991). Decreased
performance related to sleep loss and circa-
dian disruption has been implicated in some
major disasters, such as the Exxon Valdez,
Three Mile Island, and Bhopal accidents (Of-
fice of Technology Assessment, 1991). Al-
though the potential risks of sleep loss and
circadian disruption exist, they do not dimin-
ish society's need for continuous 24-h opera-
tions. Therefore, it is critical that the extent
and impact of fatigue in operational settings
be understood. Strategies and countermea-
sures should be developed and empirically
evaluated to determine approaches that will
maximize performance and alertness and
help to maintain an adequate margin of
safety.
This article describes a National Aeronau-
tics and Space Administration (NASA) pro-
HUMAN FACTORS
gram, the approach, methods, research. and
results of which are designed to address these
issues in the aviation environment. It pro-
vides an example of how, in one mode of
transportation, fatigue has been empirically
studied to provide relevant data to both the
scientific and operational communities.
NASA AMES FATIGUE
COUNTERMEASURES PROGRAM

Program Goals and Research Approach
In 1980, responding to a congressional re-
quest, NASA Ames Research Center spon-
sored a workshop that examined whether
"the circadian rhythm phenomenon, also
called jet lag," was of concern (National Aero-
nautics and Space Administration, 1980, p.
1). The workshop participants concluded that
"there is a safety problem, of uncertain mag-
nitude, due to transmeridian flying and a po-
tential problem due to fatigue in association
with various factors found in air transport
operations" (abstract). The NASA Ames Fa-
tigue/Jet Lag Program (later called the Fa-
tigue Countermeasures Program) was created
to determine the magnitude of the problem
and its operational implications. Three pro-
gram goals were established, which continue
to guide research efforts: (1) to determine the
extent offatigue, sleep loss, and circadian dis-
ruption in flight operations; (2) to determine
the impact of these factors on flight crew per-
formance; and (3) to develop and evaluate
countermeasures to mitigate the adverse ef-
fects of these factors and to maximize flight
crew performance and alertness.
The overall research approach has been to
integrate data collected from field studies
during regular flight operations with full-
mission, high-fidelity flight simulation stud-
ies and with results from controlled labora-
tory experiments. Each of these research
approaches has strengths and weaknesses

from both a scientific and an operational
FATIGUE IN AVIATION
viewpoint. Field studies may more accurately
reflect real-world conditions but are inher-
ently difficult to conduct because, for exam-
ple, research protocols must not interfere
with regular operational procedures or
safety. Also, it is virtually impossible to con-
trol all potential contributory factors in field
studies, and there are limitations to the em-
pirical measures that can be tolerated by sub-
jects in their usual work environment.
Laboratory studies provide a controlled en-
vironment for manipulating specific indepen-
dent variables and determining the outcome
in precise ways. However, attempts to gener-
alize these findings to the more complex op-
erational setting may greatly limit the oper-
ational relevance of laboratory studies that
cannot incorporate all of the potential inter-
vening variables.
Full-mission, high-fidelity flight simulation
studies provide a unique opportunity to ma-
nipulate the specifics of a trip scenario and
measure a wide range of flight variables.
However, constraints on the simulation envi-
ronment must be considered (e.g., realism,
costs). Clearly, the integration of all three ap-
proaches in a single program has important
advantages, and the program has conducted

studies using each of these research ap-
proaches to capitalize on their unique
strengths.
Measures
The program uses a diverse range of empir-
ical measures to evaluate fatigue. In any
given study, the measures are determined by
the specific hypotheses and objectives of that
particular study. The initial field studies ex-
amining the extent of fatigue, sleep loss, and
circadian disruption utilized a combination
of self-report and physiological measures.
The self-report measures included a back-
ground questionnaire and a pilot's daily
logbook.
The background questionnaire collected
June 1994-329
demographic data and information on flight
experience, sleep, meals, exercise, and gen-
eral health. It also included some standard-
ized surveys, such as that for personality
style.
The pilot's daily logbook, which is pocket-
sized, contains information about flight and
duty times, sleep quantity and quality. mood,
meals, exercise, physical symptoms, and
comments (e.g., operational events, environ-
mental factors). These data are collected a
maximum of three days prior to a trip sched-
ule, throughout a trip, and for a maximum of

four days after the trip.
The first physiological variable assessed
was core body temperature, which allowed
the determination of circadian phase and
measurement of amplitude. An ambulatory
recorder (Vitalog, Inc.) collected continuous
temperature data via a rectal thermistor,
heart rate information via chest electrode
placements, and activity detected by a mo-
tion sensor worn on the nondominant wrist.
This portable recorder collected and stored
the data for up to 10 days.
Over the years, the sophistication and
range of measures has increased. Continuous
portable recording of electroencephalo-
graphic (EEG), electro-oculographic (EOG),
and electromyographic (EMG) activity is ob-
tained using an Oxford Medilog 9200 re-
corder. This instrument provides up to eight
channels of physiological data, which are
continuously collected in analog form onto a
high-quality cassette tape. The Medilog can
collect continuous data on one tape for up to
24 h. The tape is later played back on a scan-
ning system that allows detailed analysis of
standardized sleep variables (e.g., sleep la-
tency, total sleep time, amounts of rapid eye
movement [REM] and non-REM sleep, and
awakenings).
The Medilog can provide both sleep and
wakefulness data. The EEG information can
be analyzed for EEG frequency changes (e.g.,

330-June 1994
alpha and theta activity), and the EOG data
reflect slow eye movements, which are asso-
ciated with physiological sleepiness during
wakefulness (Akerstedt and Gillberg, 1990).
In addition to these physiological mea-
sures, a recent study incorporated a vigilance
performance measure that uses reaction time
to assess sustained attention (Dinges and
Powell, 1985, 1988). This psychomotor vigi-
lance task (PVT) is a lO-min simple reaction
time test that probes central nervous system
(CNS) capability. It has been demonstrated to
be sensitive to the effects of sleep loss and
circadian disruption and does not have a sig-
nificant learning curve, unlike many other
performance tests (Dinges and Kribbs, 1991).
The PVT data are especially useful as a
metric because they can be compared with
previous findings from sleep deprivation and
sleep disorder studies. The PVT does not rep-
resent a specific flight performance variable.
Instead, it provides a measure of CNS capa-
bility, especially of sustained attention and
vigilance performance, which are important
factors in many operational settings, includ-
ing aviation.
New measures are added as study objec-
tives expand into different areas. For exam-
ple, an upcoming project will require the

measurement of noise levels with a sound
pressure meter during flight and blood oxy-
gen saturation using an oximeter during
sleep. The principal measures used were:
Background questionnaire (e.g., demographics,
personality)
Survey/questionnaire data (e.g., operational is-
sues)
Logbook subjective report (e.g., pilot's daily log-
book)
Observationallbehavioral data (e.g., cockpit ob-
server log)
Physical performance and mental functioning
tests (e.g., psychomotor vigilance task)
Long-term continuous recording of motor activ-
ity (actigraphy)
Long-term continuous recording (via Vitalog) of
physiological parameters (e.g., core body tem-
perature, heart rate)
HUMAN FACTORS
Continuous physiological recording (via Med-
ilog) of brain (EEG), eye (EOG), and muscle
(EMG) activity
Typically, one or two NASA researchers/
observers accompany volunteer pilots during
a trip to collect the measures and provide
guidance regarding the protocol.
Over the past 12 years, studies have been
conducted in a variety of aviation and con-
trolled laboratory environments and in one
full-mission flight simulation. These projects

were often labor intensive, spanned several
years from design to implementation to anal-
ysis and reporting of results, and often in-
volved collaborations. Support and collabo-
ration in the United States came from the
Federal Aviation Administration (FAA), the
National Transportation Safety Board
(NTSB), air carriers, pilots, pilot unions, and
the military. Some of the international stud-
ies involved worldwide collaborations with
research and flight operations groups from
the United Kingdom, Germany, Japan, and
other countries. References reporting the re-
sults from many of the NASA studies and
other significant program publications are
provided in the appendix.
The results of these studies have been col-
lectively organized into an extensive data-
base that encompasses data from more than
500 volunteer pilots. This database allows
unique comparisons between operational en-
vironments in which similar measures were
taken.
Another cri tical factor in successfully col-
lecting these data has been the assurance of
anonymity and confidentiality for all of the
participants. After volunteering for a study, a
pilot would receive an identification number,
and no name would ever be associated with
any of the data collected.
RESEARCH EXAMPLES
Three studies will be highlighted to provide

examples of the research approach, measures,
FATIGUE IN AVIATION
and findings and the operational relevance of
the data.
Short-Haul Commercial Operations
This study was conducted to examine the
extent of sleep loss, circadian disruption, and
fatigue engendered by flying commercial
short-haul air transport operations (flight
legs less than 8 h; see Gander, Graeber,
Foushee, Lauber, and Connell, in press). In
this study, 74 pilots from two airlines were
studied before, during, and after three- and
four-day commercial short-haul trips. All
flights took place on the East Coast of the
United States and occurred throughout the
year. Of the pilots contacted about the study,
85% agreed to participate. As a group, the pi-
lots averaged 41.3 years of age and had, on
average, 14.6 years of airline experience.
Physiological data (core body temperature
and heart rate) and motor activity were ob-
tained every 2 min with the Vitalog portable
biomedical monitor. Using the pilot's daily
logbook, subjects provided subjective ratings
of fatigue and mood every 2 h while awake
and recorded their sleep episodes and other
activities (e.g., meals, exercise, duty time). All
subjects completed a background question-

naire, and a NASA cockpit observer accom-
panied crews during trip schedules.
The specific daytime and evening trips
studied were selected so as to provide infor-
mation about the upper range of fatigue re-
ported by pilots in these operations. Common
features of the trip schedules included early
report times (i.e., for duty) and multiple flight
legs (average 5.5/day) over long duty days.
The trips averaged 10.6 h of duty per day and
involved an average of 4.5 h of flight time.
One third of the duty periods studied were
longer than 12 h. The average rest period was
12.5 h long and usually occurred progres-
sively earlier in the day across successive
trip days.
Data from the daily logbook demonstrated
June 1994-331
that during trip nights, pilots took about 12
min longer to fall asleep, slept about 1.2 h
less, and awoke about 1.4 h earlier compared
with their pretrip sleep patterns. The pilots
reported this trip sleep as lighter and poorer
(with more awakenings) than pretrip sleep.
Subjective fatigue and mood were worse dur-
ing layovers than before or after the trip or
during flights. Significant time-of-day effects
were found for fatigue, negative emotions,
and activation ratings. In the first three rat-
ings of the day following awakening, fatigue
and negative emotion ratings were low.
Thereafter they increased and reached their

highest values in the final rating prior to
sleep. Predictably, activation ratings showed
the inverse of this pattern.
On trip days, pilots consumed more caf-
feine (mean = 3.4 servings) than on pretrip
days (mean = 1.9 servings) or posttrip days
(mean = 2.7 servings), presumably to main-
tain alertness during operations. Caffeine was
consumed primarily in the early morning,
which is associated with the earlier wake-up
and duty times, and also during the midafter-
noon peak in physiological sleepiness. During
the trip schedule more alcohol (mean = 1.6
servings) was consumed than on pretrip
(mean = 0.5 servings) and posttrip (mean =
1.0 servings) days. It can be assumed that pi-
lots consumed the alcohol only after coming
off duty (presumably to unwind after a long
duty day), and in accordance with federal avi-
ation regulations. During trips, more snacks
were consumed, and they were consumed
earlier in the wake period.
For 72 pilots flying 589 legs, heart rates ob-
tained during takeoff, descent, and landing
were compared with midcruise values. The
pilot flying had greater increases in heart rate
during descent and landing than did the pilot
not flying. This increase was greater under
instrument flight rule than under visual
flight rule conditions.
This was one of the first field studies

332-June 1994
conducted by the NASA program, and it pro-
vides a unique insight into the physiological
and subjective effects of flying short-haul
commercial operations. It demonstrated that
these measures could be obtained in an oper-
ational environment without disturbing reg-
ular performance of duties. The study results
suggest several significant operational con-
siderations regarding fatigue. For example,
the data showed that the daily duty durations
were double the flight durations and that one
third of the duty periods were longer than 12
h. Findings from this study suggest that lim-
itations on duty time should be considered,
just as pilot flight times are currently limited
by federal aviation regulations. Also, the
practice of making pilots report for duty ear-
lier on successive trip days, requiring earlier
wake-up times, interferes with obtaining ad-
equate sleep. Even when the layovers were
relatively long, the circadian system would
generally inhibit falling asleep earlier, and
hence a significant amount of sleep would be
lost during trip nights. Therefore, when pos-
sible, duty on successive trip days should
begin at the same time or even begin pro-
gressively later, moving with the natural
tendency of the biological clock to extend
the day.
Finally, alcohol is known to disrupt sleep
dramatically and therefore contributes to the
poor quantity and quality of sleep obtained

on trip nights. Alternate ways to unwind after
duty and to promote sleep should be identi-
fied and offered (e.g., cognitive-behavioral re-
laxation approaches).
Long-Haul Commercial Operations
This study examined how long-haul (>8 h)
flight crews organized their sleep during a va-
riety of international trip patterns and how
duty requirements, local time, and the circa-
dian system affect the timing, quantity, and
quality of sleep (Gander, Graeber, Connell,
and Gregory, 1991). Duty requirements and
HUMAN FACTORS
local time can be viewed as externaUenviron-
mental constraints on time available for
sleep, whereas the internal circadian system
is a major physiological modulator of sleep
duration and quality.
Subjects were 29 male flight crew members
(average age = 52 yrs) flying Boeing 747 air-
craft on one of four commercial international
trip patterns. This report combined the data
from the four trip schedules. The pilot's daily
logbook was completed prior to, during, and
following the trip to collect self-reports of
duty times and of sleep timing, duration, and
quality, and so on (i.e., the same data as in the
short-haul study). Core body temperature,
heart rate, and activity using the Vitalog por-
table biomedical monitor were collected ev-
ery 2 min. The core body temperature, mea-

sured with a rectal thermistor, was used as a
marker of the underlying circadian time-
keeping system.
On average, the duty periods lasted about
10.3 h and were followed by 24.8 h oflayover.
During layovers there were generally two
sleep episodes. The average sleep/wakeful-
ness pattern was 19 h awake, 5.7 h of sleep,
7.4 h awake, and 5.8 h of sleep. Pilots gener-
ally reported that the first sleep of the layover
was of better quality and that they fell asleep
more easily and obtained deeper sleep than in
the subsequent sleep episode. As the length of
sleep increased, there was a concomitant in-
crease in sleep quality ratings. The circadian
system appeared to exert a greater influence
on the timing and duration of the first sleep
episode than on that of the second sleep of a
layover, with a preference for sleeping during
local night and/or waking up after the tem-
perature minimum. The exception was after
eastward flights that crossed five or more
time zones and produced a high accumulated
sleep debt. The time of falling asleep for the
second sleep episode was related to the
amount of sleep already obtained. It typically
occurred during local night, and the duration
FATIGUE IN AVIATION
was related to the remaining time available
before duty. The duration of both sleep epi-
sodes was longer when crew members fell

asleep earlier with respect to their circadian
temperature minimums.
Subjective reports of naps taken during
layovers were obtained from the daily log-
book. When the first sleep episode of a lay-
over was identified as a nap, it averaged 2 h
in length, was generally longer than other
naps, and followed significantly longer peri-
ods of wakefulness. These first naps typically
occurred in response to acute sleep loss asso-
ciated with overnight eastward flights or
westward flights crossing five time zones or
more. Other naps occurred just prior to the
next duty period and effectively shortened
the length of continuous wakefulness.
Crew members also reported in their log-
books the occurrence of naps on the flight
deck. (In-flight rest on the flight deck is not
sanctioned under current federal regula-
tions.) The average duration of the naps re-
ported on the flight deck was 46 min (range,
10--130min). Research observers accompany-
ing the crews also noted naps not reported in
the logbooks. Data combining the research
observers' notes and logbook data suggest
that, on average, 11% of flight crew members
were taking the opportunity to nap when con-
ditions permitted. The data do not indicate
whether these were planned naps or occurred
spontaneously in response to sleep loss and
circadian disruption.
This study provides unique insights into
the physiological and subjective effects of fly-

ing long-haul commercial operations. The in-
formation is scientifically provocative and
can be translated into operationally relevant
considerations. The following are some of the
scientific considerations that emerge from
the results.
The flight schedules pushed the sleep/wake
cycle into a period (25.7 h) different from that
of the circadian system, though the two sys-
June 1994-333
terns did not become completely uncoupled.
Although the circadian system continued to
influence the timing and duration of sleep ep-
isodes, it was unable to resynchronize and
quickly adapt to the rapid, multiple time-
zone shifts. It is clear that a variety of exter-
nal/environmental factors (e.g., light, activ-
ity, social cues) interact with internal!
physiological factors to affect sleep timing,
duration, and quality. The operational rele-
vance of the data is easy to discern. For ex-
ample, current flight and duty time regula-
tions are intended to ensure that reasonable
minimum rest periods are available for flight
crews. However, this study demonstrated
that in commercial long-haul flight sched-
ules, there are physiologically and environ-
mentally determined preferred sleep times
within a layover, and therefore the time
available for sleep may be less than the off-
duty time available.
Planned Cockpit Rest

As indicated from the results of the previ-
ous study, long-haul flight operations involve
sleep loss and circadian disruption. Anec-
dotal, observational, and self-report sources
(e.g., those from the previous study) indicate
that sleep does occur on the flight deck, de-
spite federal regulations forbidding in-flight
rest. It is unclear from available data how
often these naps are planned and how often
they occur spontaneously in response to sleep
loss and circadian disruption. In consider-
ation of the available information regarding
rest on the flight deck, the first test of an op-
erational fatigue countermeasure was con-
ducted. A NASA/FAAstudy examined the ef-
fectiveness of a planned cockpit rest period to
maintain and/or improve subsequent perfor-
mance and alertness in long-haul, non aug-
mented international flights (nonaugmented
= only primary crew required; augmented =
extra crew needed when over certain flying
times; Rosekind, et ai., in press).
334-June 1994
A regularly scheduled 12-day, eight-leg trip
that involved multiple trans-Pacific crossings
was selected for study. The flight legs aver-
aged just over 9 h and were followed by about
25 h of layover. Prior to, during, and after the
12-day trip, crew members completed pilot's
daily logbooks, documenting sleep and duty
times and other activities, and wore acti-

graphs on their nondominant wrist. (An acti-
graph collects noninvasive activity data from
a motion sensor, providing an estimate of an
individual's 24-h rest/activity pattern; Cole,
Kripke, Gruen, Mullaney, and Gillin, 1992.)
The middle four legs of the trip schedule were
studied intensively. EEG and EOG activity
was continuously monitored during these
flight legs using the Medilog recorder. An ex-
tensive scientific literature demonstrates that
self-reports of sleep (e.g., time to fall asleep,
total sleep time) do not accurately reflect
physiological activity (Carskadon et aI., 1976;
Rosekind and Schwartz, 1988). Therefore it
was critical to document the amount of phys-
iological sleep obtained. Continuous EEG
and EOG recordings taken during the awake
period were used to assess physiological
sleepiness. The PVT was used as a measure of
vigilance performance and sustained atten-
tion. Pilots also gave self-report ratings of
alertness and mood at predetermined times
throughout the flight. Two NASA researchers
traveled with the crews to implement the
procedures and collect data.
The three-person Boeing 747 volunteer
crew members were randomly assigned to ei-
ther a rest group or no-rest group. Each rest
group member (12 subjects) had a 40-min rest
opportunity during the low-workload portion
of flights over water. Crew members rested
one at a time on a prearranged rotation while
the other two crew members maintained the
flight. The no-rest group (9 subjects) had a
40-min control period identified when they

were instructed to continue their regular
flight activities. Specific safety and proce-
HUMAN FACTORS
dural guidelines were used during the study.
The first question was whether pilots
would be able to sleep during a planned rest
opportunity in their cockpit seat. The rest
group slept during 93% of the rest opportuni-
ties. On average, they fell asleep in 5.6 mins
and slept for about 26 min. Whether or not a
benefit was associated with this sleep was de-
termined by examining the vigilance perfor-
mance measure and indicators of physiologi-
cal sleepiness. As expected, the no-rest group
showed performance decrements (i.e., in-
creased reaction times and variability on the
PVT) at the end of flights compared with the
beginning of flights, on night flights versus
day flights, and on the fourth study leg com-
pared with the first study leg. The rest group,
however, demonstrated positive effects of the
brief nap by maintaining consistently good
performance at the end of flights, on night
flights, and on the fourth study leg.
Physiological sleepiness was examined by
evaluating the subtle EEG and EOG changes
that indicate state lability. Previous research
has demonstrated that physiological sleepi-
ness is associated with the occurrence of EEG
alpha or theta waves and/or EOG slow eye
movements. These physiological events are
associated with decreased performance (Ak-

erstedt, 1992; Akerstedt and Gillberg, 1990).
Microevents (brief sleep events) indicative of
physiological sleepiness (the occurrence of
EEG alpha or theta waves and/or EOG slow
eye movements) lasting 5 s or longer were
identified during the last 90 min of flight,
even during descent and landing, in both
study groups. Overall, the no-rest group had
microevents (mean = 6.37) indicative of
physiological sleepiness at a rate twice that of
the rest group (mean = 2.90).
The brief nap appeared to act as an acute
in-flight operational safety valve and did not
affect the cumulative sleep debt observed in
85% of the crew members. The rest group
members were usually able to sleep during
FATIGUE IN AVIATION
the rest opportunity, and this nap was asso-
ciated with improved performance and alert-
ness compared with the no-rest control
group. This was the first empirical test of a
fatigue countermeasure conducted in an op-
erational aviation setting that combined
physiological, performance, and subjective
measures.
Not only are the results scientifically inter-
esting, but they can be transferred directly
into operational considerations regarding
planned rest. Based partly on the results of
this NASA/FAA study, an industry/govern-

ment working group has drafted an advisory
circular for review by the FAA's Aviation
Rulemaking Advisory Committee. The circu-
lar outlines specific guidelines for the devel-
opment and implementation of a program for
controlled rest on the flight deck. It should be
noted that controlled rest is only one in-flight
countermeasure and is not the panacea for all
of the sleep loss and circadian disruption en-
gendered by long-haul flight operations
(Rosekind, Gander, and Dinges, 1991).
Current Activities and Future Directions
In 1991, the name of the program was
changed to the Fatigue Countermeasures Pro-
gram to emphasize the development and
evaluation of countermeasures. One area of
intense activity is the analysis and writing of
a variety of scientific and operational publi-
cations to transfer the information acquired
over the past 12 years to the scientific and
operational communities. Another project in-
volves a study of the onboard crew rest facil-
ities on long-haul aircraft. Bunks are avail-
able for pilots when a flight is augmented
(extra crew onboard) and the flight length ex-
tends beyond that allowed for a single crew.
This study will examine the quantity and
quality of sleep obtained in onboard crew rest
facilities, the factors that promote or inter-
fere with sleep, and the effects on subsequent
performance and alertness.
June 1994-335

Another major project is the development
and implementation of an education and
training module entitled" Alertness Manage-
ment in Flight Operations" (Rosekind, Gan-
der, and Connell, in press). It provides infor-
mation on physiological sources of fatigue
and how flight operations affect these factors
and makes recommendations for fatigue
countermeasures. The module is intended for
any interested party in the aviation commu-
nity, including pilots, air carrier managers,
schedulers, flight attendants, and federal pol-
icymakers. Potential areas for future program
activity include the development of an expert
scheduling system that incorporates known
scientific and physiological data, an exami-
nation of fatigue in regional airline opera-
tions, and further development and evalua-
tion of countermeasures.
FUTURE CONSIDERATIONS FOR
FATIGUE RESEARCH IN
OPERATIONAL SETTINGS
As issues of safety and health continue to be
raised regarding fatigue in operational envi-
ronments, more research will be required to
address them. Major questions raised in
many operational settings include, What is
safe? How long is too long to drive, fly, or
operate a train? How many night shifts in a
row is too many? How long is too long for a
shift period? How many consecutive days can
be worked safely? How long does it take to
recover after an extended duty or shift pe-

riod? How long does it take to recover after
several consecutive duty or shift periods? Do
individuals need one night of sleep or two
nights? How could naps be used to improve
the situation? How should one define recov-
ery: by physiological adaptation, perfor-
mance, or waking sleepiness? After starting a
new shift schedule, how long will it take to
physiologically adapt? What are the effects of
changing from a night shift schedule to a reg-
ular daytime schedule on the weekends and
336--June 1994
then back to night shifts? How should current
hours-of-service regulations be evaluated? If
they should be changed, how should they be
adjusted? On what should such changes be
based?
Each of the issues raised suggests a wide
range of research activities. They also repre-
sent the concerns faced by the 20 million
Americans currently working altered or non-
standard shifts. The point of these questions
is that the scientific research will be most
useful if it is integrated with operational con-
cerns.
There are other considerations as well. Al-
though generic research will help to set a sci-
entific foundation for addressing these oper-
ational questions, it is also critical to
understand the specific requirements of dif-

ferent settings. The type of shift schedule,
task demands, timing of critical tasks, and
other factors such as meal availability and
break schedules present different challenges.
Another area to examine is the tremendous
individual variation that exists in response to
sleep loss, circadian disruption, and perfor-
mance changes. Different work and corporate
cultures will have disparate attitudes, knowl-
edge, and concern about these issues, and this
can affect work performance and satisfaction.
Also, there is very little information regard-
ing the long-term (e.g., months or years) ef-
fects of sleep loss and circadian disruption on
safety, performance, productivi ty, and
health. Another major area is the develop-
ment and empirical evaluation of counter-
measures. There may be a wide variety of so-
lutions that will maximize alertness and
performance-for instance, those involving
physiological strategies, display design, or
schedule design. This area creates tremen-
dous potential for evaluating new technolo-
gies and strategies.
One valuable approach is to coordinate and
integrate resources and expertise among re-
searchers, federal agencies, policymakers,
HUMAN FACTORS
and so on to maximize the potential effect of
any given research project or operational im-
plementation of findings. It is often critical
that scientists be allowed access to the sub-
jects' operational environment or, at mini-

mum, to fully understand how things work in
a particular setting. The close coordination of
scientific and operational efforts can also lay
the foundation for a direct application of pos-
itive findings.
Clearly there is a lot of important work to
be done. The NASA Ames Fatigue Counter-
measures Program is presented as a model of
one approach to addressing some of the issues
raised. It is not the only approach possible
but is presented to stimulate empirical re-
search in applied operational settings.
Human factors research can playa pivotal
role in the scientific investigation of these is-
sues and in the application of the findings to
specific operational questions. As long as 24-h
operations are required to maintain societal
needs, the effects of sleep loss and circadian
disruption must be considerations in any op-
erational setting. How these factors affect the
physiological and performance capabilities of
the human operator will be critical to job
safety, performance, and productivity.
ACKNOWLEDGMENTS
We wish to acknowledge the significant contributions of
the following: John Lauber, Curt Graeber, Clay Foushee,
and Charles Billings; William Reynard and Linda Connell
at NASA Ames; Key Dismukes for a thorough and con-
structive manuscript review; David Dinges, Institute for
Pennsylvania Hospital/University of Pennsylvania School
of Medicine, for ongoing collaboration; and the FAA staff,
volunteer pilots. air carriers, and unions, who have been

critical to the success of program activities,
APPENDIX: FATIGUE
COUNTERMEASURES PROGRAM
REFERENCES (PARTIALLISTING)
Dinges, D. F., and Graeber, R. C. (1989, October). Crew fa-
tigue monitoring. Flight Safety Digest, (Suppl.), 65-75.
Dinges. D. F.•Connell. L J" Rosekind, M. R., Gillen, K. A.,
Kribbs, N. B., and Graeber, R. C. (1991). Effects of
cockpit naps and 24-h layovers on sleep debt in long-
haul transmeridian flight crews. Sleep Research, 20,
406. (Abstract)
FATIGUE IN AVIATION
Dinges, D. F., Graeber, R. C., Connell, L. J. Rosekind,
M. R., and Powell, J. W. (1990). Fatigue-related reac-
tion time performance in long-haul flight crews. Sleep
Research, 19, 117. (Abstract)
Dinges, D. F., Rosekind, M. R., Connell, L. J., Graeber,
R. C., and Gillen, K. A. (1992). Eastbound night flights
vs. westbound day flights: Directionally dependent ef-
fects on flight crew layover sleep. Sleep Research, 21,
118. (Abstract)
Foushee, H. C., Lauber, J. K., Baetge, M. M. and Acombe,
D. B. (1986). Crew factors in flight operations: 111. The
operational significance of exposure to short-haul air
transport operations (Tech. Memorandum 88322). Mof-
fett Field, CA: National Aeronautics and Space Admin-
istration.

Gander, P. H., Barnes, R., Gregory, K. B., Connell, L. J.,
Miller, D. L., and Graeber, R. C. (in press). Crew factors
in flight operations: Vll. Psychophysiological responses
to helicopter operations (Tech. Memorandum). Moffett
Field, CA: National Aeronautics and Space Administra-
tion.
Gander, P. H., Connell, L. 1., and Graeber, R. C. (1986).
Masking of the circadian rhythms of heart rate and
core body temperature by the rest-activity cycle in
man. Journal of Biological Rhythms Research, 1, 119-
135.
Gander, P. H., and Graeber, R. C. (1987). Sleep in pilots
flying short-haul commercial schedules. Ergonomics,
3D, 1365-1377.
Gander, P. H., Gregory, K. 8., Connell, L. J., Miller, D. L.,
and Graeber, R. C. (in press). Crew factors in flight op-
erations: Vll. Psychophysiological responses to overnight
cargo operations (Tech. Memorandum). Moffett Field,
CA: National Aeronautics and Space Administration.
Gander, P. H., Kronauer, R., and Graeber, R. C. (1985).
Phase-shifting two coupled circadian pacemakers: Im-
plications for jetlag. American Journal of Physiology,
249, R704-R719.
Gander, P. H., McDonald, J. A., Montgomery, J. c., and
Paulin. M. G. (1991). Adaptation of sleep and circadian
rhythms to the Antarctic summer: A question of zeit-
geber strength? Aviation, Space and Environmental
Medicine, 62, 1019-1025.
Gander, P. H., Myrhe, G., Graeber, R. C., Anderson, H. T.,

and Lauber, J. K. (1989). Adjustment of sleep and the
circadian temperature rhythm after flights across nine
time zones. Aviation, Space and Environmental Medi-
cine, 60, 733-743.
Gander, P. H., Nguyen, D., Rosekind, M. R., and Connell,
L. J. (1993). Age, circadian rhythms, and sleep loss in
flight crews. Aviation, Space and Environmental Medi-
cine, 64, 189-195.
Gander, P. H., and Same!. A. (1991). Shiftwork in space:
Bright light as a chronobiologic countermeasure (SAE
Tech. Paper 911496). Salem, MA: Society of Automo-
tive Engineers.
Graeber, R. C. (1982). Alterations in performance follow-
ing rapid transmeridian flight. In F. M. Brown and
R. C. Graeber (Eds.), Rhythmic aspects of behavior (pp.
172-212). Hillsdale, NJ: Erlbaum.
Graeber, R. C. (Ed.). (1986). Crew factors in flight opera-
tions: IV. Sleep and wakefulness in international air-
crews (Tech. Memorandum 88231). Moffett Field, CA:
National Aeronautics and Space Administration.
Graeber, R. C. (1988). Aircrew fatigue and circadian rhyth-
micity. In E. Wiener and D. Nagel (Eds.), Human fac-
tors in aviation (pp. 305-344). New York: Academic.
June 1994-337
Graeber, R. C. (1989). Jet lag and sleep disruption. In M. H.
Kryger, T. Roth, and W. C. Dement (Eds.), Principles
and practice of sleep medicine. Philadelphia: Harcourt
Brace Jovanovich.

Graeber, R. C., Foushee, C., Gander, P. R., and Noga, G.
(1985). Circadian rhythmicity and fatigue in flight op-
erations. Journal of Occupational and Environmental
Health, 7(Suppl.), 122-129.
Graeber, R. C., Foushee, C., and Lauber, J. (1984). Dimen-
sions of flight crew performance decrements: Method-
ological implications for field research. In J. Cullen
and J. Siegrist (Eds.), Breakdown in human adaptation
to stress (pp. 584-605). Boston: Martinus Nijhoff.
Kurosaki, Y., Sasaki, M., Takahashi, T., Mori, A., Spinwe-
ber, C. L., and Graeber, R. C. (1987). Flight crew sleep
after multiple layover polar flights. Sleep Research, 16,
565. (Abstract)
Monk, T., Moline, M., and Graeber, R. C. (1988). Inducing
jet lag in the laboratory: Patterns of adjustment to an
acute shift in routine. Aviation, Space and Environmen-
tal Medicine, 59, 703-710.
Rosekind, M. R. (1992). Pilots. In M. A. Carskadon (Ed.),
The encyclopedia of sleep and dreaming. New York: Mac-
millan.
Rosekind, M. R., Dinges, D. F., Gregory, K. B., Gillen,
K. A., Smith, R. M., Powell, J. W., and Miller, D. L.
(1993). Estimating nap sleep in operational settings: A
comparison of actigraphy vs. ambulatory polysomnog-
raphy. Sleep Research, 22, 380. (Abstract)
Rosekind, M. R., Gander, P. H., Miller, D. L., Gregory,
K. B., McNally, K. L., Smith, R. M., and Lebacqz, J. V.
(1993). NASA Ames fatigue countermeasures program.
FAA Aviation Safety Journal, 3(1), 20-25.

Same!. A., and Gander, P. H. (Eds.). (1991). Light as a chro-
nobiologic countermeasure for long-duration space oper-
ations (Tech. Memorandum 103874). Moffett Field, CA:
National Aeronautics and Space Administration.
REFERENCES
Akerstedt, T. (1991). Sleepiness at work: Effects of irregu-
lar work hours. In T. Monk (Ed.), Sleep, sleepiness, and
performance(pp. 129-152). Chichester, England: Wiley.
Akerstedt, T. (1992). Work hours and continuous monitor-
ing of sleepiness. In R. J. Broughton and R. D. Ogilvie
(Eds.), Sleep, arousal, and performance (pp. 63-72). Bos-
ton: Birkhauser.
Akerstedt, T., and Gillberg, M. (1990). Subjective and ob-
jective sleepiness in the active individual. International
Journal of Neuroscience, 52, 29-37.
Bonnett, M. H. (1985). The effect of sleep disruption on
performance, sleep, and mood. Sleep, 8, 11-19.
Broughton, R. J., and Ogilvie, R. D. (Eds.). (1992). Sleep,
arousal, and performance. Boston: Birkhiiuser.
Carskadon, M. A., and Dement, W. C. (1982). Nocturnal de-
terminants of daytime sleepiness. Sleep, 5, S73-S81.
Carskadon, M. A., Dement, W. C., Mitler, M. M., Guillemi-
nault, C., Zarcone, V. P., and Spiegel, R. (1976). Self-
reports versus sleep laboratory findings in 122 drug-
free subjects with complaints of chronic insomnia.
American Journal of Psychiatry, 133, 1382-1388.
Cole, R. J., Kripke, D. F., Gruen, W., Mullaney, D. J., and

Gillin, J. C. (1992). Automatic sleep/wake identification
from wrist activity. Sleep, IS, 461-469.
Dinges, D. F. (1989). The influence of the human circadian
time keeping system on sleep. In M. H. Kryger, T. Roth,
338-June 1994
and W. C. Dement (Eds.), Principles and practice of sleep
medicine (pp. 153-162). Philadelphia: W.B. Saunders.
Dinges, D. F., and Kribbs, N. B. (1991). Performing while
sleepy: Effects of experimentally induced sleepiness. In
T. Monk (Ed.), Sleep, sleepiness, and performance (pp.
97-128). Chichester, England: Wiley.
Dinges, D. F., and Powell, J. W. (1985). Microcomputer
analyses of performance on a portable, simple visual
RT task during sustained operations. Behavior Re-
search Methods, Instruments and Computers, 17, 652-
655.
Dinges, D. F., and Powell, J. W. (l988). Sleepiness is more
than lapsing. Sleep Research, 17, 84. (Abstract)
Gander, P. H., Graeber, R. C., Connell, L. J., and Gregory,
K. B. (l991). Crew factors in flight operations: VIll. Fac-
tors influencing sleep timing and subjective sleep quality
in commercial long-haul flight crews (Tech. Memoran-
dum 103852). Moffett Field, CA: National Aeronautics
and Space Administration.
Gander, P. H., Graeber, R. C., Foushee, H. C., Lauber, J. K.,
and Connell, L. 1. (in press). Crew factors in flight oper-

ations: ll. Psychophysiological responses to short-haul
air transport operations (Tech. Memorandum 88321).
Moffett Field, CA: National Aeronautics and Space Ad-
ministration.
Mistlberger, R., and Rusak, B. (I 989}. Mechanisms and
models of the circadian timekeeping system. In M. H.
and practice of sleep medicine (pp. 141-152). Philade]-
phia: W.B. Saunders.
Monk, T. H. (1989). Circadian rhythms in subjective acti-
vation, mood, and performance efficiency. In M. H.
and practice of sleep medicine (pp. ]63-172). Philadel-
phia: W.B. Saunders.
National Aeronautics and Space Administration. (1980).
HUMAN FACTORS
Pilot fatigue and circadian desynchronis (Report of a
workshop held in San Francisco, California, on August
26-28, 1980; Tech. Memorandum 81275). Moffett Field,
CA: Author.
Office of Technology Assessment. (1991, September). Bio-
logical rhythms: Implications for the worker (OTA-BA-
463). Washington, DC: U.S. Government Printing Of-
fice.
Rosekind, M. R., Gander, P. H., and Connell, L. J. (in
press). Crew factors in flight operations: X. Alertness
management in flight operations (Tech. Memorandum).
Moffett Field, CA: National Aeronautics and Space Ad-
ministration.

Rosekind, M. R., Gander, P. H., and Dinges, D. F. (1991).
Alertness management in flight operations: Strategic nap-
ping (SAE Tech. Paper Series 912138). Salem, MA: So-
ciety of Automotive Engineers.
Rosekind, M. R.o Graeber, R. C., Dinges, D. F., Connell,
L. J., Rountree, M. S., Spinweber, C. L., and Gillen,
K. A. (in press). Crew factors in flight operations: IX.
Effects of planned cockpit rest on crew performance and
alertness in long-haul operations (Tech. Memorandum
103884). Moffett Field, CA: National Aeronautics and
Space Administration.
Rosekind, M. R., and Schwartz, G. E. (l988). The percep-
tion of sleep and wakefulness: I. Accuracy and cer-
tainty of subjective judgments. Sleep Research, 17, 89.
(Abstract)
Roth, T., Roehrs, T.A., Carskadon, M. A., and Dement,
W. C. (1989). Daytime sleepiness and alertness. In
M. H. Kryger, T. Roth, and W. C. Dement (Eds.), Prin-
ciples and practice of sleep medicine (pp. ]4-23). Phila-
delphia: W.B. Saunders.
Date received: January 25, 1993
Date accepted: July 22, 1993
INTRODUCTION
With the rapid advancement of technology,
complex work systems have evolved in which op-
erators must adapt their decision making and per-

formance in the face of dynamic, ever-changing
environments, concurrent task demands, time
pressure, and tactical constraints (Moray, 1997;
Sheridan, 2002). The assessment and prediction
of the mental workload associated with operating
such complex systems has long been recognized
as an important issue (e.g., Gopher & Donchin,
1986; Moray, 1979). Mental workload – or just
workload – is the general term used to describe
the mental cost of accomplishing task require-
ments (Hart & Wickens, 1990; Wickens, 1992).
Workload varies as a function of task demands
placed on the human operator and the capacity of
the operator to meet those demands (Gopher &
Donchin, 1986; Hopkin, 1995). High levels of
workload occur when task demands exceed oper-
ator capacity.
Research efforts in complex work systems such
as piloting (e.g., Wilson, 2002), unmanned aerial
vehicle control (e.g., Dixon, Wickens, & Chang,
2005), anesthesiology (e.g., Leedal & Smith, 2005),
railway signaling (e.g., Pickup et al., 2005), and
automobile driving (e.g., Recarte & Nunes, 2003)
have focused on identifying factors that influence
mental workload and techniques for measuring it.
In contrast, in the current paper we develop a the-
oretical model of operator strategic behavior and
workload management, within the context of en
route air traffic control (ATC), through which task-
related workload can be predicted within com-
plex work systems. The model we present takes
into account the changing task priorities and man-
agement of resources by operators as well as the
feedback that operators receive in response to their

input. In the next section, we explain the nature of
Modeling and Predicting Mental Workload in En Route Air
Traffic Control: Critical Review and Broader Implications
Shayne Loft, Penelope Sanderson, Andrew Neal, and Martijn
Mooij, The University of
Queensland, Brisbane, Australia
Objective: We perform a critical review of research on mental
workload in en route
air traffic control (ATC). We present a model of operator
strategic behavior and work-
load management through which workload can be predicted
within ATC and other
complex work systems. Background: Air traffic volume is
increasing worldwide. If
air traffic management organizations are to meet future demand
safely, better models
of controller workload are needed. Method: We present the
theoretical model and then
review investigations of how effectively traffic factors, airspace
factors, and opera-
tional constraints predict controller workload. Results:
Although task demand has a
strong relationship with workload, evidence suggests that the
relationship depends
on the capacity of the controllers to select priorities, manage
their cognitive resources,
and regulate their own performance. We review research on
strategies employed by
controllers to minimize the control activity and information-
processing requirements
of control tasks. Conclusion: Controller workload will not be
effectively modeled until
controllers’ strategies for regulating the cognitive impact of

task demand have been
modeled. Application: Actual and potential applications of our
conclusions include
a reorientation of workload modeling in complex work systems
to capture the dynam-
ic and adaptive nature of the operator’s work. Models based
around workload regula-
tion may be more useful in helping management organizations
adapt to future control
regimens in complex work systems.
Address correspondence to Shayne Loft, ARC Key Centre for
Human Factors and Applied Cognitive Psychology, University
of Queensland, Brisbane 4072, Australia; [email protected]
HUMAN FACTORS, Vol. 49, No. 3, June 2007, pp. 376–399.
DOI 10.1518/001872007X197017. Copyright © 2007, Human
Factors and Ergonomics Society. All rights reserved.
PREDICTING MENTAL WORKLOAD 377
the work performed by en route air traffic con-
trollers (ATCos) and provide an overview of our
approach.
OVERVIEW
Controlled airspace is divided into sectors. An
en route sector is a region of airspace that is typ-
ically situated at least 30 miles (~48 km) from an
airport for which an associated ATCo has respon-
sibility. ATCos have to accept aircraft into their
sector; check aircraft; issue instructions, clear-
ances, and advice to pilots; and hand aircraft off
to adjacent sectors or to airports. The radar screen

displays characteristics of the sector (e.g., bound-
aries and airways), the spatial position of aircraft,
and vital flight information (identifiers, altitude,
speed, flight destination). When the aircraft leaves
the airspace assigned to the ATCo, control of the
aircraft passes on to ATCo controlling the next
sector (or to the tower ATCo). As is typical in
many real-world complex systems, this environ-
ment imposes multiple concurrent demands on the
operator. As Gronlund, Ohrt, Dougherty, Perry,
and Manning (1998) described it, “In the en route
air traffic control environment (involving the high-
speed and high-altitude cruise between takeoff
and landing), the system that confronts the air traf-
fic controller comprises a large number of aircraft
coming from a variety of directions, at diverse
speeds and altitudes, heading to different desti-
nations. Like most complex, dynamic systems,
this one cannot be periodically halted while the
controller takes a brief respite” (p. 263).
ATCos have two main goals. The primary goal
is to ensure that aircraft under jurisdiction adhere
to International Civil Aviation Organization
(ICAO) mandated separation standards. For ex-
ample, one of the most common separation stan-
dards requires that aircraft under radar control be
separated by at least 1,000 feet vertically (2,000
feet above 29,000 feet, unless reduced vertical
separation minima apply) and 5 nautical miles
horizontally. The secondary goal is to ensure that
aircraft reach their destinations in an orderly and
expeditious manner. These goals require the ATCo
to perform a variety of tasks, including monitoring
air traffic, anticipating loss of separation (i.e., con-
flicts) between aircraft, and intervening to resolve

conflicts and minimize disruption to flow. (For an
extensive compilation of the tasks and goals of en
route control, see Rodgers & Drechsler, 1993.)
Total world airline scheduled passenger traffic
in terms of passenger-kilometers is projected to
grow at an annual average rate of 4.4% over the
period 2002 to 2015, according to forecasts pre-
pared by the ICAO (2004). In the United States
alone, the number of aircraft handled by ATC
centers is expected to increase from 46.2 million
in 2004 to more than 60.2 million in 2016 (Feder-
al Aviation Administration, 2005). To accommo-
date predicted traffic growth there is a need to
increase en route airspace capacity through the
introduction of new air traffic management sys-
tems (e.g., free flight) or the adaptation of existing
airspace designs (e.g., sector boundaries), con-
troller tools (e.g., conflict resolution), and proce-
dures (e.g., reduced separation minima). The
consensus among research and operational com-
munities is that it is important to understand the
factors that drive mental workload if they are to
improve airspace capacity (Christien, Benkouar,
Chaboud, & Loubieres, 2003; Majumdar, Ochieng,
McAuley, Lenzi, & Lepadatu, 2004).
Most research has focused on identifying char-
acteristics of the air traffic picture that create task
demand for ATCos (e.g., Grossberg, 1989; Kirwan,
Scaife, & Kennedy, 2001; Manning, Mills, Fox,
& Pfleiderer, 2001). These characteristics include
the number of aircraft in transition though a sec-
tor, the number of aircraft changing altitude, and
the number of potential conflicts. Several research
groups have attempted to predict mental workload

on a moment-to-moment basis by using linear
combinations of task demand factors as predic-
tors. The resulting sets of task demand factors are
known as dynamic density metrics. Studies have
shown that the dynamic density of the airspace at
a given moment accounts for approximately half
the variance in workload at that point in time
(e.g., Kopardekar & Magyarits, 2003; Laudeman,
Shelden, Branstrom, & Brasil, 1998). In psycho-
logical terms, this represents relatively strong pre-
diction. In practical terms, however, a significant
proportion of variance remains unaccounted for.
Furthermore, one goal of workload modeling is
to allow ATC providers to predict workload levels
ahead of time in order to allow them to put work-
load management strategies in place. For example,
this may include splitting a sector or introducing
flow restrictions. However, to date, dynamic den-
sity metrics have been unable to accurately pre-
dict ATCo workload ahead of time (Kopardekar
& Magyarits, 2003; Majumdar & Ochieng, 2002;
378 June 2007 – Human Factors
Masalonis, Callaham, & Wanke, 2003). Many re-
searchers argue that these limitations stem from
the fact that there is no simple linear relationship
between task demand and workload (e.g., Athènes,
Averty, Puechmorel, Delahaye, & Collet, 2002;
Chatterji & Sridhar, 2001). Moreover, these re-
searchers view workload as an emergent property
of the complex interaction between the ATCo and
the air traffic situation, rather than as a simple
outcome of task demand inputs at a single point

in time.
The approach to predicting ATCo mental work-
load that is presented in the current paper is in line
with these views. According to Sperandio (1971),
workload is not something imposed upon a pas-
sive ATCo but, rather, is something the ATCo
actively manages. He proposed a model in which
changes in strategy (primarily resource manage-
ment) allow ATCos to regulate how task demands
are transformed into workload, thus keeping
workload within acceptable limits. In a paper that
deserves to be better known, Rouse, Edwards, and
Hammer (1993) took a similar view, modeling
workload as a feedback control process driven by
subjective mental workload. Several current re-
search groups agree with Sperandio’s (1971) view
that a relationship between task demand and work-
load can be better understood by considering how
ATCos use strategies to manage their resources
and regulate their workload (Athènes et al., 2002;
Averty, Collet, Dittmar, Vernet-Maury, & Athènes,
2004; Cullen, 1999; Hilburn, 2004; Histon &
Hansman, 2002; Majumdar et al., 2004). Another
key aspect of Sperandio’s (1971) approach is that
the effect of ATCo control actions on the system
is fed back to the ATCo, such that future task de-
mands are actively regulated by the ATCo (Pawlak,
Brinton, Crouch, & Lancaster, 1996).
In the current paper, we present a model of
mental workload that puts ATCos in the loop with
air traffic events, reacting to the consequences of
their own proactive behavior. Without denying the
validity of the task demand approach, we argue
that the link between task demand and workload

is largely connected to the manner in which ATCos
manage their resources. We begin by outlining a
general model of workload in ATC and contrast
it with previous approaches. We then review task
demand research in ATC. A characteristic of this
research that limits its interpretability is the ex-
tremely large list of methods and task demand
factors that have been reported. (For exhaustive
reviews, see Hilburn, 2004; and Mogford, Gutt-
man, Morrow, & Kopardekar, 1995.) The model
presented in this paper provides a framework for
integrating this literature and studying its poten-
tial strengths and shortcomings. We then turn our
attention to the smaller body of research that has
focused on ATCo control strategies (e.g., Amaldi
& Leroux; 1995; Histon & Hansman, 2002) and
to task models that have been built to simulate the
performance of ATCos (e.g., Callantine, 2002;
Leiden, Kopardekar, & Green, 2003). Finally, we
review models in which researchers have attempt-
ed to integrate task demands with human perfor-
mance models in order to predict workload (e.g.,
Averty et al., 2004; Cullen, 1999).
MENTAL WORKLOAD MODELING
ARCHITECTURES
In this section we outline different modeling
architectures that have been used to understand
ATCo mental workload. At the end, we present the
architecture that guides our review of the litera-
ture that follows.
Common Architectures

A prevalent model in ATC mental workload re-
search is that different properties, or task demands,
of the air traffic situation will pose problems of
different levels of complexity to the ATCo and,
depending on the ATCo’s skill, experience, and
strategy, will produce different levels of subjective
workload. This is effectively an open-loop model,
as shown in Figure 1. Variants of this general open-
loop architecture have been used to model sources
of ATCo workload. For example, Hilburn and
Jorna’s (2001) model shows system factors com-
bining to create task demand, which, in accordance
with operator factors such as skill, strategy, and
experience, will lead to some degree of workload.
Similarly, Mogford et al.’s (1995) model depicts
a relationship between source factors (objective
complexity, air traffic patterns, sector character-
istics) and workload being mediated by quality
of equipment, individual differences, and ATCo
strategies. Both these models acknowledge that
ATCo strategy can influence workload. Neverthe-
less, researchers using these models as frame-
works have tended to focus on whether individual
aspects of the ATC environment affect workload.
We argue that the tendency to seek input-output
relations (“does increasing the number of aircraft
increase workload?” or “how does strategy influ-
ence the impact of the number of aircraft on work-
load?”) fails to take into account fully the goals
and management of resources by the ATCo and
feedback that the ATCo receives from the system

in response to his or her input.
Sperandio’s Architecture
It is interesting to contrast the aforementioned
models with that of Sperandio (1971), shown in
Figure 2. Again, task demand is on the left and
mental workload on the right. Consistent with the
previous two models, Sperandio (1971) proposed
that ATCo strategy is an intervening variable be-
tween task demand and the work achieved and that
the ATCo selects strategies to keep mental work-
load within acceptable limits. However, in contrast
to the models mentioned in the previous para-
graph, the Sperandio (1971) model includes two
feedback control loops. First, variation in mental
workload resulting from work methods has,
through feedback, a regulating effect on the choice
of work methods (Feedback Loop 1). Second, the
work method used in response to perceived task
demands regulates the task demand encountered
in the future (Feedback Loop 2). Sperandio (1971)
emphasized that it is the change in workload, not
the change in task demand, that explains the
change in strategy. The change in strategy changes
what information is extracted from the airspace
and thus affects workload. The relationship among
task demand, strategy, and workload is adaptive
and so can be shown only in a feedback control
diagram, as shown in Figure 2. This position is
consistent with that of Rouse et al. (1993) but is
in contrast to much subsequent research in which
mental workload is predicted from objectively
measured characteristics of the airspace, tuned by
strategy and other factors.

A Systems Approach to Modeling Mental
Workload
The basis of our approach is that the ATCo is
in a continuous relationship with a dynamic world
and is an adaptive element in that world (Athènes
et al., 2002; Pawlak et al., 1996; Sperandio, 1971).
As a result, mental workload cannot be a function
solely of task demands; it is also a function of the
strategy the ATCo uses to manage traffic and
whether the strategy, once invoked, has provided
Air traffic
factors
Mediating
factors
Mental
workload
Task
demandAir traffic
factors
Mediating
factors
Mental
workload
Task
demand
Figure 1. Generic form of an open-loop model of mental

workload implicitly adopted in many studies of ATCo men-
tal workload.
Input:
Work
assigned
Choice of
work method
Output:
Work
done
Feedback loop 1
Feedback loop 2
Task demand
Input: Output:
Feedback loop 1
Feedback loop 2
Mental
workload
Task demand
Figure 2. Sperandio’s (1971) closed-loop model of ATCo
mental workload in which actual work done influences
choice of work methods and amount of work accepted. Task

demand is equivalent to work to be done. Variation of
operator’s strategies and regulating effects on workload, J. C.
Sperandio, Ergonomics, (1971), Taylor & Francis Ltd,
adapted by permission of the publisher (Taylor & Francis Ltd,
http://www.tandf.co.uk/journals).
a comfortable level of control over task demands.
In the next section, we work through the ATC lit-
erature with the help of the model that is shown in
Figure 3. In an adaptation of Sperandio’s (1971)
model, our model represents two control loops that
govern ATCo activity: first, the management of
workload by the internal reorganization of prior-
ities leading to a different strategy; and second,
the management of workload by explicit control
of the airspace. The work of the ATCo occupies
the shaded part in the centre of Figure 3, where-
as the world that the ATCo controls is shown in
the perimeter.
The model in Figure 3 simplifies the world
that the ATCo controls. It shows two aircraft,
which is the minimum needed to indicate that the
ATCo is concerned with managing relationships
among aircraft rather than controlling single air-
craft in isolation. One aircraft is at the top of Fig-
ure 3 and the other at the bottom. Each aircraft
receives instructions from the ATCo (see plus
signs [+] on links into adder symbols at top right
and bottom right of Figure 3). The pilot considers
the difference between the instruction and the air-

craft’s current flight profile (see minus sign [–]
on feedback loops coming into rightmost adder
symbols) and makes an appropriate adjustment
to the aircraft’s flight profile. The aircraft’s new
flight profile is combined with the flight profile of
all other aircraft (see two + signs entering adder
at left) and becomes the task demand fed back to
the ATCo. The ATCo can take action to change fu-
ture task demand fed back through the system
(e.g., by accepting aircraft early or by putting air-
craft in a holding pattern) or he or she can change
future task demand with cognitive strategies (e.g.,
by considering a set of aircraft as one for purpos-
es of control).
The ATCo remains aware of work to be done
by monitoring present task demands (see left of
Figure 3). The work to be done is translated into
actual work done through the control activities
performed. As Figure 3 suggests, a set of control
activities can be classified as a strategy. A strategy
can be described as a specific class of air traffic
management that achieves one or more objectives
(e.g., safety, orderliness, expeditiousness) with a
certain investment of time and effort. Selection
A/C 2
+
+
PilotA/C 1
Pilot A/C 2

Control
activities/
strategies
Prioritization
Meta
cognition
+
-
-
demands
A/C 2
PilotA/C 1
PilotA/C 2
ATCo
Control
activities/
strategies
Prioritization
Meta
cognition
Feedforward Feedback

Task
Task
demands
Work to
be done
Actual work
done
+
+
-
Figure 3. Model of ATCo activity in which strategy is
controlled through feedback and feedforward information about
mental workload. Work to be done represents how objective
task demands are mentally represented by the ATCo.
among strategies is driven by the relative priority
of the ATCo’s objectives as the work to be done
evolves over time (Kallus, Van Damme, & Ditt-
man, 1999; Kirwan & Flynn, 2002; Niessen,
Eyferth, & Bierwagen, 1999). The strategy chosen
will lead to a certain quality of control over the
situation, which will often reflect subjective time
pressure. For example, the contextual control mod-
el (COCOM) of Hollnagel (2002) and Hollnagel
and Woods (2005) distinguishes strategic, tactical,

opportunistic, and scrambled control as the result
of an operator’s subjective judgment of the rela-
tionship between time available for action (Ta) and
the time required to evaluate the situation (Te),
select a response (Ts), and perform the response
(Tr). In their COCOM model, strategic and tacti-
cal control are mostly proactive, whereas oppor-
tunistic and scrambled control are mostly reactive.
An ATCo will work to achieve strategic control
and avoid scrambled control as much as possible.
Figure 3 indicates that prioritization drives
control activities/strategies. Prioritization refers
to the professional set of values that guide control
of air movements at any point, such as safety,
orderliness, and expeditiousness. Prioritization,
in turn, is driven by metacognitive factors such as
awareness of time available to perform tasks,
anticipation of future difficulties, and the ATCo’s
knowledge of his or her capacity. However, we do
not assume that this is a conscious process. In-
stead, prioritization is a consideration satisfied
directly or indirectly through control action. The
strategy may be a learned response that is not open
to introspection, and it may not require explicit
consideration of safety, orderliness, or expedi-
tiousness. However, this does not remove the fact
that strategies will always reflect some balance of
priority among safety, orderliness, and expedi-
tiousness. The selection of strategy will be logi-
cally guided by the appropriate priority even if the
priority is not consciously accessed.
Focusing in particular on time, Schmidt (1978)
predicted ATCo mental workload with a queuing
theory model based on the relationship between

the frequency of observable tasks that require de-
cisions/actions and the time required to make these
decisions/actions. More recently, Hendy, Liao,
and Milgram (1997) used a simulated ATC task and
modeled overall workload as a univariate function
of time pressure. Time pressure, in turn, was mod-
eled as the ratio of time available to time required,
also expressible as the ratio of the information-
processing rate demanded by a task and the max-
imum information-processing capacity of the
ATCo. Performance data were well matched by
the model. Most recently, Rantanen and Levinthal
(2005) demonstrated that an ATCo’s time to first
intervention in resolving conflicts was faster when
the ratio of time to act to the duration of a window
of opportunity for action was small – in other
words, when there was less discretionary time.
The dashed lines within the shaded ATCo part
of Figure 3 indicate that metacognition can be
influenced by both feedforward and feedback sig-
nals. Looking at feedforward (at left), the ATCo
may be aware that a large number of aircraft are
about to enter the sector and thus adjusts his or her
strategy for handling traffic already on frequency.
Looking at feedback (at right), the ATCo may be
aware that the quality of actual work done may
have been compromised by time pressure and thus
adjust priorities toward achieving safety at the
possible expense of expeditiousness (e.g., by rear-
ranging the trajectories of aircraft in order to min-
imize monitoring and coordination requirements).
The adder to the right of metacognition in Figure
3 shows that the ATCo may notice differences
between the work to be done and the actual work

that is getting done, which will also trigger a meta-
cognitive response. For example, the ATCo may
realize that a heavy communication load is mak-
ing him or her fall behind in dealing with work to
be done, yet an even heavier communication load
is anticipated. A rearrangement of priorities may
offer a control strategy that has a less intense com-
munications load. In these ways the model in Fig-
ure 3 represents the fact that ATCos behave and
react to the consequences of their behavior – in
respect both to the perceived discrepancy between
current goals and system state and to the ATCo’s
understanding of his or her own capacity – and
that these processes drive mental workload.
In the following sections we review research
on the relationships between task demand and
mental workload and between operator capacity
and mental workload. Then we review research
that combines the two in ways that are at least par-
tially consistent with the model in Figure 3.
TASK DEMANDS AND MENTAL
WORKLOAD
Research concerning the relationship between
task demand and mental workload has had a long
history, dating back more than 40 years (Arad,
1964; Couluris & Schmidt, 1973; Davis, Dana-
her, & Fischl, 1963; Hurst & Rose, 1978; Schmidt,
1976). This research has focused on uncovering

properties of the air traffic environment that con-
tribute to cognitive complexity and, via tuning fac-
tors such as skill, strategy and experience, result
in workload (see Figure 1). The task demand liter-
ature, in its own right, provides a solid basis from
which to model the complexity of task demands
(work to be done) in the ATC system. Researchers
have expended great effort in developing predic-
tive models based on task demand, and the fact
that these models have been able to account for
significant variance in ATCo workload warrants
a state-of-the-art review and synthesis. We main-
tain, however, that focusing only on task demand
overlooks the reciprocal interactions presented in
Figure 3. Subsequently, we examine whether the
task demand literature explicitly or implicitly mod-
els these aspects of control, and we outline con-
tradictions in the literature that result from the
failure to take a systems view.
Our review of task demand research is sum-
marized in Table 1. The material is drawn from
government and contractor technical reports, op-
erational reviews, journal articles, and book chap-
ters. Much material originates in the United States
and Europe, primarily from the Federal Aviation
Administration, the National Aeronautics and
Space Administration (NASA), and the European
Organisation for the Safety of Air Navigation
(EUROCONTROL). During the review process,
we found that several aspects of task demand
research limited its interpretability. Two valuable
surveys of research relating to task demand (Hil-
burn, 2004; Mogford et al., 1995) do not explicit-
ly address these issues. These constraints, and how
we dealt with them, will be briefly discussed next.

First, systematic comparison among studies
was complicated by the wide variety of research
methodologies reported. As presented in Table 1,
these methodologies include knowledge elicita-
tion techniques such as verbal protocol analysis
(e.g., Pawlak et al., 1996), experiments in which
researchers made predictions a priori about how
mental workload will vary with systematic manip-
ulation of task demand (e.g., Boag, Neal, Loft, &
Halford, 2006), and correlational studies in which
researchers extracted values for task demand fac-
tors from flight data and correlated these values
with workload on a post hoc basis (e.g., Kopar-
dekar & Magyarits, 2003). A second characteris-
tic of task demand research that limits cross-study
comparison is the wide variety of workload mea-
sures used. An evaluation of the advantages and
disadvantages of each approach is beyond the
scope of this paper (see Farmer & Brownson,
2003; Hilburn & Jorna, 2001). However, Table 1
categorizes studies according to the workload cri-
terion employed. As is evident from Table 1, the
measurement of workload is far from uniform.
Table 1 reveals that most studies have focused
on identifying traffic factors. Traffic factors re-
flect the instantaneous distribution of air traffic in
a sector in terms of both the number of aircraft and
the complexity of their relationships. However,
ATCos can actively regulate the mental workload
associated with traffic factors by using economi-
cal control strategies. Researchers such as Hilburn
(2004) and Histon and Hansman (2002) identified
two further factors that influence the choice and

effectiveness of ATCo control strategies and which
are generally independent from traffic factors.
Airspace factors reflect the underlying structural
properties of the airspace (e.g., number of cross-
ing altitude profiles). Airspace factors constrain
the relationship between traffic factors and work-
load by shaping the evolution of air traffic and
creating predictable air traffic patterns that can be
exploited by ATCos. Operational constraints re-
flect operational requirements (e.g., restrictions of
available airspace) that place restrictions on ATCo
strategy and control action.
Traffic Factors Predicting Mental
Workload
Task demand research has typically focused on
explicit properties of the distribution of aircraft
that predict mental workload and are computed in
real time using radar track data or derivations
thereof. Of all the traffic factors, the aircraft count,
or the number of aircraft under control, is the most
powerful predictor of workload (e.g., Hurst &
Rose, 1978; Kopardekar & Magyarits, 2003; Man-
ning et al., 2001). High aircraft count leads to an
increase in workload because it increases the mon-
itoring, communication and coordination required
to handle aircraft in a safe, orderly, and expedi-
tious manner.
Density factors are derivatives of traffic count
that measure the horizontal and vertical distances
between aircraft and the way in which these dis-
tances change with time. Various measures of

traffic density have been developed, including the
average horizontal (or lateral) separation distances
between aircraft (Chatterji & Sridhar, 2001) and
aircraft counts divided by sector volumes (Kopar-
dekar & Magyarits, 2003). These density measures
assume, perhaps erroneously, that all aircraft in
close proximity are noticed by the ATCo and there-
fore exert some influence on mental workload. In
reality, the influence of aircraft density on work-
load will depend on what the aircraft are doing in
relation to each other – for example, whether they
are converging or diverging. For this reason, den-
sity factors based on the minimum separations be-
tween pairs of aircraft are more sensitive (Chatterji
& Sridhar, 2001). Close future minimum separa-
tion between an aircraft pair will undoubtedly cap-
ture ATCo attention because of the likelihood of
a separation violation, reducing the ATCo’s capac-
ity to attend to other control tasks.
Although traffic count and density adequately
reflect the number of routine aircraft-associated
tasks that an ATCo has to perform within a certain
time frame, the complexity of air traffic is impor-
tant in determining the difficulty of the tasks or
events handled and thus the resulting mental
workload. ATCos report that they can handle rel-
atively large volumes of traffic if the aircraft are
flying on regular routes and the flow is orderly
(e.g., Amaldi & Leroux, 1995; Mogford et al.,
1995). In contrast, small volumes of traffic can
lead to overload if aircraft interact in complex
ways (Kallus, Van Damme, & Dittman, 1999;

Mogford et al., 1995). Complexity factors fall
into two categories. The first are commonly re-
ferred to as aircraft transition factors and capture
changes in an aircraft’s state in any of the three
axes of altitude (e.g., Lamoureux, 1999), speed
(e.g., Kopardekar & Magyarits, 2003), or heading
(e.g., Laudeman et al., 1998). Performance mix
of aircraft is also an important aircraft transition
factor (Schaefer, Meckiff, Magill, Pirard, & Aligne,
2001). In order to calculate the minimum separa-
tion between two aircraft in altitude transition,
the ATCo needs to know how fast the aircraft will
climb (or descend) relative to each other. Pre-
sumably, this job would be made harder by in-
creased variability in performance profiles.
The second set of complexity factors relates to
the number and nature of potential conflicts with-
in a sector. Potential conflicts emerge from the
combination of density and transition factors pre-
sent at any time. The influence of potential con-
flicts on mental workload depends on their spe-
cific properties. For example, Boag et al. (2006)
developed a “transitions metric” for assessing the
difficulty of judging whether a pair of aircraft will
be in lateral and vertical conflict at the same time.
If two aircraft are on converging flight paths, and
are both maintaining the same level, then the
ATCo simply needs to assess whether they will
violate the lateral separation standard. This prob-
lem is relatively simple because the ATCo need
consider only one transition (the transition into lat-
eral conflict). However, if the aircraft are chang-
ing levels, then the ATCo must assess when the
aircraft will violate and regain separation in one

dimension and also whether the aircraft will be in
conflict in the other dimension at the same time.
Up to four transitions (into and out of lateral and
vertical conflict) may be possible. The Boag et al.
(2006) transitions metric accounted for signifi-
cant amounts of variance in ATCos’ ratings of
complexity and workload. Imminent violations
of separation also increase workload (Chatterji &
Sridhar, 2001). The time available for an ATCo to
detect and respond to a potential conflict affects
how difficult the conflict is to resolve. A conflict
that develops quickly gives the ATCo only a lim-
ited time to act, creating significant time pressure.
Conflicts in close proximity to sector boundaries
(e.g., Pawlak et al., 1996) and/or high numbers of
surrounding aircraft (e.g., Kopardekar & Magyar-
its, 2003) also constrain how conflicts can be
resolved, reducing the number of options for
maneuvering.
Combining Task Demand Factors Into
Dynamic Density Metrics
Several research groups (Kopardekar & Mag-
yarits, 2003; Laudeman et al., 1998; Masalonis
et al., 2003) have used regression models to iden-
tify sets of factors that best predict mental work-
load and have created algorithms by weighting
these different factors according to their predictive
power. The resulting algorithms are commonly
referred to as dynamic density metrics. Dynamic
density has been defined as “the collective effort
of all factors, or variables, that contribute to sector-
level air traffic control complexity or difficulty at
any point in time” (Kopardekar & Magyarits,
2003, p. 1). The Laudeman et al. (1998) metric is

perhaps the best known, describing dynamic den-
sity as the sum of the density of traffic weighted
by the number of changes in speed, heading, and
384
TA
B
LE
1
:
S
tu
d
ie
s
P
re
d
ic
ti
n
g
A
T

C
o
M
e
n
ta
l W
o
rk
lo
ad
:
M
e
th
o
d
s
o
f
M
e
as
u
re

m
e
n
t
an
d
S
u
m
m
ar
ie
s
o
f
R
e
su
lt
s
A
u
th
o
rs
M

e
th
o
d
M
e
as
u
re
s
Ta
sk
D
e
m
an
d
F
ac
to
rs
S
ys
te
m

s
V
ie
w
A
ra
d
(
1
9
6
4
)
C
o
rr
e
la
ti
o
n
al
A
T
C
o

a
ct
iv
it
y
Tr
af
fi
c,
a
ir
sp
ac
e
,
an
d
S
e
ct
o
r
b
o
u

n
d
ar
ie
s
al
ig
n
e
d
w
it
h
s
ta
n
d
ar
d
fl
o
w
s
re
d

u
ce
o
p
e
ra
ti
o
n
al
c
o
n
st
ra
in
ts
th
e
t
im
e
p
re

ss
u
re
o
f
co
n
tr
o
l a
ct
iv
it
ie
s.
B
o
ag
e
t
al
.
E
xp

e
ri
m
e
n
t:
p
ar
t-
ta
sk
M
W
r
at
in
g
s
Tr
af
fi
c
T
h
e
n

u
m
b
e
r
o
f
co
n
fl
ic
t
b
o
u
n
d
ar
y
tr
an
si
ti
o
n
s

in
cr
e
as
e
s
(2
0
0
6
)
si
m
u
la
ti
o
n
C
o
m
p
le
xi
ty

r
at
in
g
s
M
W
b
e
ca
u
se
A
T
C
o
s
ca
n
n
o
t
re
d

u
ce
in
fo
rm
at
io
n
s
e
t.
B
u
ck
le
y,
D
e
B
ar
ys
h
e
,
E
xp

e
ri
m
e
n
t:
f
u
ll
ta
sk
A
T
C
o
a
ct
iv
it
y
Tr
af
fi
c
an

d
a
ir
sp
ac
e
A
ir
sp
ac
e
f
ac
to
rs
in
te
ra
ct
w
it
h
t
ra
ffi

c
fa
ct
o
rs
in
d
e
te
r-
H
it
ch
n
e
r,
&
K
o
h
n
si
m
u
la

ti
o
n
m
in
in
g
M
W
;
A
T
C
o
s
u
se
a
ir
sp
ac
e
s
tr

u
ct
u
re
t
o
s
im
p
lif
y
(1
9
8
3
)
ai
r
tr
af
fi
c.
C
h
at

te
rj
i &
S
ri
d
h
ar
C
o
rr
e
la
ti
o
n
al
:
d
yn
am
ic
M
W
r

at
in
g
s
Tr
af
fi
c
M
o
d
e
l i
n
co
rp
o
ra
te
s
co
g
n
it

iv
e
a
sp
e
ct
s
(e
.g
.,
t
im
e
p
re
s-
(2
0
0
1
)
d
e
n
si
ty
su

re
)
in
to
c
h
o
ic
e
o
f
tr
af
fi
c
co
m
p
le
xi
ty
f
ac
to
rs

.
C
o
u
lu
ri
s
&
S
ch
m
id
t
E
xp
e
ri
m
e
n
t:
F
u
ll
ta

sk
A
T
C
o
a
ct
iv
it
y
A
ir
sp
ac
e
U
si
n
g
A
T
C
o
a

ct
iv
it
y
as
m
e
as
u
re
o
f
M
W
f
ai
ls
t
o
t
ak
e
in
to

(1
9
7
3
)
si
m
u
la
ti
o
n
ac
co
u
n
t
th
e
in
te
n
t
o
f

co
n
tr
o
l a
ct
io
n
.
D
av
is
e
t
al
.
(1
9
6
3
)
E
xp
e
ri

m
e
n
t:
F
u
ll
ta
sk
A
T
C
o
a
ct
iv
it
y
Tr
af
fi
c
U
si
n

g
A
T
C
o
a
ct
iv
it
y
as
m
e
as
u
re
o
f
M
W
f
ai
ls
t

o
t
ak
e
in
to
si
m
u
la
ti
o
n
ac
co
u
n
t
th
e
in
te
n
t

o
f
co
n
tr
o
l a
ct
io
n
.
G
al
st
e
r
e
t
al
.
(2
0
0
1
)

E
xp
e
ri
m
e
n
t:
P
ar
t-
ta
sk
M
W
r
at
in
g
s
Tr
af
fi
c
D

e
la
ys
in
c
o
n
fl
ic
t
d
e
te
ct
io
n
im
p
o
se
c
o
n
st
ra

in
ts
o
n
si
m
u
la
ti
o
n
S
e
co
n
d
ar
y
ta
sk
co

n
fl
ic
t
re
so
lu
ti
o
n
.
H
is
to
n
&
H
an
sm
an
K
n
o
w
le

d
g
e
e
lic
it
at
io
n
:
liv
e
C
o
m
p
le
xi
ty
r
at
in
g
s
Tr

af
fi
c,
a
ir
sp
ac
e
,
an
d
A
ir
sp
ac
e
f
ac
to
rs
in
te
ra
ct
w

it
h
t
ra
ffi
c
fa
ct
o
rs
in
d
e
te
r-
(2
0
0
2
)
o
b
se
rv

at
io
n
a
n
d
in
te
rv
ie
w
o
p
e
ra
ti
o
n
al
c
o
n
st

ra
in
ts
m
in
in
g
M
W
;
A
T
C
o
s
u
se
a
ir
sp
ac
e
s
tr
u

ct
u
re
t
o
s
im
p
lif
y
ai
r
tr
af
fi
c.
H
u
rs
t
&
R
o
se
(

1
9
7
8
)
C
o
rr
e
la
ti
o
n
al
:
fu
ll
ta
sk
A
T
C
o
a
ct
iv

it
y
Tr
af
fi
c
U
si
n
g
A
T
C
o
a
ct
iv
it
y
as
m
e
as
u

re
o
f
M
W
f
ai
ls
t
o
t
ak
e
si
m
u
la
ti
o
n
in
to

a
cc
o
u
n
t
th
e
in
te
n
t
o
f
co
n
tr
o
l a
ct
io
n
.
K
ir
w

an
e
t
al
.
(2
0
0
1
)
K
n
o
w
le
d
g
e
e
lic
it
at
io
n
:

C
o
m
p
le
xi
ty
r
at
in
g
s
Tr
af
fi
c,
a
ir
sp
ac
e
,
an
d
A

T
C
o
s
u
se
a
ir
sp
ac
e
s
tr
u
ct
u
re
t
o
s
im
p
lif
y

ai
r
tr
af
fi
c.
g
ro
u
p
ju
d
g
m
e
n
t
o
p
e
ra
ti
o
n
al

c
o
n
st
ra
in
ts
K
o
p
ar
d
e
ka
r
&
C
o
rr
e
la
ti
o
n

al
:
u
n
ifi
e
d
C
o
m
p
le
xi
ty
r
at
in
g
s
Tr
af
fi
c
an
d

a
ir
sp
ac
e
A
ir
sp
ac
e
f
ac
to
rs
p
la
y
an
im
p
o
rt
an
t

ro
le
in
p
re
d
ic
ti
n
g
M
ag
ya
ri
ts
(
2
0
0
3
)
d
yn
am
ic

d
e
n
si
ty
M
W
a
cr
o
ss
s
e
ct
o
rs
.
La
m
o
u
re
u
x

(1
9
9
9
)
E
xp
e
ri
m
e
n
t:
p
ar
t-
ta
sk
M
W
r
at
in
g
s
Tr

af
fi
c
M
W
is
li
n
ke
d
t
o
t
h
e
m
e
n
ta
l c
al
cu
la
ti
o

n
s
an
d
p
ro
je
ct
io
n
s
si
m
u
la
ti
o
n
in
vo
lv
e
d

in
m
an
ag
in
g
g
ro
u
p
s
o
f
A
/C
.
La
u
d
e
m
an
e
t
al

.
K
n
o
w
le
d
g
e
e
lic
it
at
io
n
:
A
T
C
o
a
ct
iv
it

y
Tr
af
fi
c
A
ck
n
o
w
le
d
g
e
d
t
h
at
v
ar
ia
ti
o
n
in

t
h
e
in
te
n
t
o
f
A
T
C
o
(1
9
8
8
)
in
te
rv
ie
w
co

n
tr
o
l a
ct
io
n
c
an
in
fl
u
e
n
ce
t
h
e
r
e
la
ti
o
n
sh
ip

b
e
tw
e
e
n
C
o
rr
e
la
ti
o
n
al
:
d
yn
am
ic
A
T
C
o

a
ct
iv
it
y
an
d
M
W
.
d
e
n
si
ty
385
M
an
n
in
g
e

t
al
.
C
o
rr
e
la
ti
o
n
al
:
A
T
C
o
a
ct
iv
it
y
M
W

r
at
in
g
s
Tr
af
fi
c
A
T
C
o
a
ct
iv
it
y
d
o
e
s
n
o
t

si
g
n
ifi
ca
n
tl
y
p
re
d
ic
t
M
W
o
ve
r
(2
0
0
1
)
fr
o

m
r
o
u
ti
n
e
fl
ig
h
t
d
at
a
an
d
a
b
o
ve
t
ra
ffi
c
co

u
n
t;
f
ai
ls
t
o
t
ak
e
in
to
a
cc
o
u
n
t
th
e
in
te
n
t

o
f
co
n
tr
o
l a
ct
io
n
.
M
an
n
in
g
,
M
ill
s,
F
o
x,

C
o
rr
e
la
ti
o
n
al
:
A
T
C
o
a
ct
iv
it
y
M
W
r
at
in
g

s
Tr
af
fi
c
N
o
.
an
d
t
yp
e
o
f
A
T
C
o
c
o
m
m
u
n

ic
at
io
n
s
d
o
n
o
t
si
g
n
ifi
-
P
fl
e
id
e
re
r,
&
fr
o

m
r
o
u
ti
n
e
fl
ig
h
t
d
at
a
ca
n
tl
y
p
re
d
ic
t
M
W

;
fa
ils
t
o
t
ak
e
in
to
a
cc
o
u
n
t
th
e
in
te
n
t
M
o
g

ilk
a
(2
0
0
2
)
o
f
A
T
C
o
c
o
m
m
u
n
ic
at
io
n
s.
M

as
al
o
n
is
e
t
al
.
K
n
o
w
le
d
g
e
e
lic
it
at
io
n
:

C
o
m
p
le
xi
ty
r
at
in
g
s
Tr
af
fi
c,
a
ir
sp
ac
e
,
an
d
M

e
tr
ic
u
n
ab
le
t
o
a
cc
u
ra
te
ly
p
re
d
ic
t
M
W
a
h

e
ad
o
f
(2
0
0
3
)
in
te
rv
ie
w
o
p
e
ra
ti
o
n
al
c
o
n
st

ra
in
ts
ti
m
e
;
fa
ils
t
o
t
ak
e
in
to
a
cc
o
u
n
t
th
e
in

fl
u
e
n
ce
o
f
A
T
C
o
C
o
rr
e
la
ti
o
n
al
:
d
yn
am

ic
co
n
tr
o
l s
tr
at
e
g
y/
ac
ti
vi
ty
o
n
f
u
tu
re
t
as
k

d
e
m
an
d
t
h
at
is
d
e
n
si
ty
fe
d
b
ac
k
th
ro
u
g
h
s

ys
te
m
.
M
e
tz
g
e
r
&
P
ar
as
u
r-
E
xp
e
ri
m
e
n
t:
p

ar
t-
ta
sk
M
W
r
at
in
g
s
Tr
af
fi
c
D
e
la
ys
in
c
o
n
fl
ic

t
d
e
te
ct
io
n
im
p
o
se
c
o
n
st
ra
in
ts
o
n
am
an
(

2
0
0
1
)
si
m
u
la
ti
o
n
S
e
co
n
d
ar
y
ta
sk
co
n
fl
ic

t
re
so
lu
ti
o
n
.
M
o
g
fo
rd
e
t
al
.
K
n
o
w
le
d
g

e
e
lic
it
at
io
n
:
C
o
m
p
le
xi
ty
r
at
in
g
s
Tr
af
fi
c,

a
ir
sp
ac
e
,
an
d
A
ir
sp
ac
e
f
ac
to
rs
a
n
d
o
p
e
ra

ti
o
n
al
c
o
n
st
ra
in
ts
p
la
y
a
ro
le
(1
9
9
3
)
ra
ti

n
g
a
n
d
r
an
ki
n
g
o
p
e
ra
ti
o
n
al
c
o
n
st
ra
in

ts
in
s
h
ap
in
g
t
h
e
r
e
la
ti
o
n
sh
ip
b
e
tw
e
e
n
t
ra

ffi
c
fa
ct
o
rs
a
n
d
M
W
.
P
aw
la
k
e
t
al
.
(1
9
9
6

)
K
n
o
w
le
d
g
e
e
lic
it
at
io
n
:
M
W
r
at
in
g
s
Tr

af
fi
c,
a
ir
sp
ac
e
,
an
d
E
m
p
h
as
is
o
n
f
ac
to
rs
t
h

at
im
p
ac
t
o
n
t
h
e
c
o
g
n
it
iv
e
ve
rb
al
p
ro
to
co
l t

e
ch
n
iq
u
e
C
o
m
p
le
xi
ty
r
at
in
g
s
o
p
e
ra
ti
o

n
al
c
o
n
st
ra
in
ts
ac
ti
vi
ty
o
f
A
T
C
o
s,
a
s
o
p

p
o
se
d
t
o
b
e
h
av
io
ra
l i
n
d
ic
at
o
rs
(e
.g
.,
A
T

C
o
a
ct
iv
it
y)
.
S
ch
ae
fe
r
e
t
al
.
K
n
o
w
le
d
g

a
ir
sp
ac
e
,
an
d
A
ir
sp
ac
e
f
ac
to
rs
in
te
ra
ct
w
it
h
t

ra
ffi
c
fa
ct
o
rs
in
d
e
te
r-
(2
0
0
1
)
in
te
rv
ie
w
o
p
e
ra

ti
o
n
al
c
o
n
st
ra
in
ts
m
in
in
g
M
W
.
S
te
in
(
1
9
8

5
)
E
xp
e
ri
m
e
n
t:
fu
ll
ta
sk
M
W
r
at
in
g
s
Tr
af
fi
c

N
o
.
o
f
h
an
d
o
ff
s
(in
b
o
u
n
d
a
n
d
o
u
tb
o
u
n
d

)
p
re
d
ic
t
M
W
;
si
m
u
la
ti
o
n
h
o
w
e
ve
r,
h
ig
h

ly
c
o
rr
e
la
te
d
w
it
h
A
/C
c
o
u
n
t.
W
yn
d
e
m
e
re

,
In
c.
K
n
o
w
le
d
g
e
e
lic
it
at
io
n
:
M
W
r
at
in
g

s
Tr
af
fi
c
an
d
a
ir
sp
ac
e
M
e
tr
ic
in
cl
u
d
e
d
s
e
ve

ra
l a
ir
sp
ac
e
f
ac
to
rs
.
(1
9
9
6
)
cr
it
ic
al
d
e
ci
si
o
n

C
o
rr
e
la
ti
o
n
al
:
d
yn
am
ic
d
e
n
si
ty
N
o
te
.

M
W
=
m
e
n
ta
l w
o
rk
lo
ad
;
A
/C
=
a
ir
cr
af
t.
altitude; the proximity of aircraft; and the time un-

til predicted conflicts. This metric accounted for
22% of the variance in ATCo activity (a proxy for
workload) not predicted by aircraft count. Kopar-
dekar and Magyarits (2003) incorporated 23 fac-
tors from four published dynamic density metrics
to form a composite metric. This unified dynam-
ic density accounted for 39% of the variance in
ATCo complexity ratings, significantly more than
aircraft count alone.
Airspace Factors and Operational
Constraints Predicting Mental Workload
Airspace factors and operational constraints
are key contributors to ATCo task demand and
mental workload (Histon & Hansman, 2002; Kir-
wan et al., 2001). Airspace factors refer to the un-
derlying structural properties of the airspace,
whereas operational constraints refer to temporary
variations in operational conditions within the air-
space. Airspace factors constrain the relationship
between traffic factors and workload by shaping
the evolution of air traffic and creating predictable
air traffic patterns. Knowledge of these patterns
lets the ATCo use information-processing stra-
tegies that simplify air traffic management. Fur-
thermore, temporary variations in operational
conditions, such as communications limitations
(Mogford, Murphy, Yastrop, Guttman, & Roske-
Hofstrand, 1993), can restrict ATCo control action
and strategy.
Studies examining airspace factors have found
that the size of a sector can influence the mental
workload imposed by traffic factors (Arad, 1964;
Histon & Hansman, 2002). On the one hand, a

larger sector size will typically increase the num-
ber of aircraft in the sector and the number of po-
tential events that require attention. On the other
hand, events evolve faster in smaller sectors, and
limited space in a sector can reduce the options
for conflict resolution. Increased number of avail-
able flight levels can reduce workload because
they allow ATCos to maintain separation using
vertical separation (Histon & Hansman, 2002;
Kirwan et al., 2001). Traffic events occurring close
to the outside of sector boundaries are important
further determinants of workload (Couluris &
Schmidt, 1973; Histon & Hansman, 2002) be-
cause they can cause the ATCo’s “area of regard”
to be greater than the official dimensions of the
sector. Aircraft events occurring outside sector
boundaries require attention because they can
affect aircraft currently in the sector, increasing
the complexity of coordination (handoffs, point
outs) with adjacent ATCos (e.g., Kirwan et al.,
2001; Mogford et al., 1993). However, the pres-
ence of well-defined ingress and egress points
(Histon & Hansman, 2002) lets ATCos anticipate
problems, thus reducing workload.
The number, orientation, and complexity of
standard flows also influence the mental workload
imposed by traffic factors (Histon & Hansman,
2002; Schaefer et al., 2001). Standard flows are
aircraft flow patterns that emerge from underly-
ing airway structure, standardized procedures,
and other regular constraints such as ingress and
egress points. ATCos use their knowledge of stan-
dard flows to create important structure-based
abstractions that simplify the management of air

traffic (Histon & Hansman, 2002). For example,
ATCos can simplify the search process involved in
conflict detection by focusing on known crossing
points and/or known crossing altitude profiles be-
tween standard flows (Histon & Hansman, 2002;
Pawlak et al., 1996). Knowledge of these “hot
spots” can reduce the workload associated with
managing air traffic (Amaldi & Leroux, 1995;
Kallus, Van Damme, & Dittman, 1999).
Several operational constraints place restric-
tions on ATCo control action. For example, re-
strictions on available airspace can result from
convective weather, activation of special-use air-
space, or aircraft in holding patterns (Kirwan et
al., 2001; Mogford et al., 1993). Restrictions on
available airspace increase the likelihood of sep-
aration violations. In addition, restricted airspace
requires more precisely planned conflict resolution
strategies. Procedural restrictions, such as miles-
in-trail spacing, can also constrain traffic flow and
conflict resolution strategies (Histon & Hansman,
2002; Pawlak et al., 1996).
Because of practical constraints, weightings for
task demand factors are typically validated against
one or only a few sectors (Histon & Hansman,
2002), limiting the variation observed in airspace
factors and operational constraints. As a result, dy-
namic density metrics developed using specific
sectors perform less effectively when extended to
other sectors. However, two research groups have
recently incorporated airspace factors into their
metrics (Kopardekar & Magyarits, 2003; Masa-
lonis et al., 2003). For example, the Kopardekar
and Magyarits (2003) unified metric was devel-

oped across four sectors, and the metric performs
differently across them. Nevertheless, comparisons
across the different sectors revealed the contri-
bution of airspace factors. Factors with significant
predictive value included the altitude level of the
sector (high vs. low), the structure of the airspace,
and the size of the sector. It seems that if research-
ers wish to predict mental workload across sectors,
then airspace factors and operational constraints
will be important because they mediate the effect
of traffic factors on workload. However, if re-
searchers wish to predict workload within sectors,
then the airspace factors and operational con-
straints become less important.
Summary and Assessment
Research focusing on task demand factors has
shown that task demand accounts for a significant
amount of variance in mental workload. We sought
to develop our model of workload by investigating
how different task demand factors might affect
ATCos’ selection of strategies for control and thus
affect workload. However, this was made difficult
by the lack of research examining how airspace
structure and aircraft configuration relate to the
selection of strategies for control, as depicted by
the model in Figure 3. We argue that a significant
limitation of the task demand approach is that it
views the ATCo as a passive recipient of task de-
mand. It does not explicitly take into account the
fact that ATCos can actively take steps that change
task demand, so as to keep workload at an accept-

HUMAN FACTORS, 1994,36(2),327-338Fatigue in Operational Se.docx

HUMAN FACTORS, 1994,36(2),327-338Fatigue in Operational Se.docx

Recommended

Recommended

More Related Content

Similar to HUMAN FACTORS, 1994,36(2),327-338Fatigue in Operational Se.docx

Similar to HUMAN FACTORS, 1994,36(2),327-338Fatigue in Operational Se.docx (20)

More from wellesleyterresa

More from wellesleyterresa (20)

Recently uploaded

Recently uploaded (20)

HUMAN FACTORS, 1994,36(2),327-338Fatigue in Operational Se.docx