Human Factors in

Aviation Safety:

A View From the Left Seat

Captain Robert Sumwalt

Chairman
Human Factors and Training Group
Air Line Pilots Association, International
Hmm…
The airplanes
are not the
only things
that have
changed
Line Pilot Priorities

1. Safety
2. Passenger Comfort
3. Schedule
4. Efficiency
 Accident rate for each
generation of aircraft

10
Transport Jet Aircraft
Hull loss per million
9
departures
8
7
6
5
4 1st generation
3
2 All aircraft
1 3rd generation 2nd generation
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39
Years Of Operation
 Human Factors
• “Human factors is a multidisciplinary
field which examines the relationship
between humans and their behavior
and activities to understand the
capabilities and limitations in order to
reduce errors and optimize performance
within a system.”
– Edwards and Hawkins
Human
Factors
Design

Engineering
Flight Test
Airworthiness
Certification
 Flight Test Engineering

Human
Factors

Airworthiness Design
Certification
ALPA Human Factors and
Training Focused Agenda
• Improving Flight Crew Monitoring Skills
• Threat and Error Management Training
• Improved Procedures and Training for Use
of Flight Deck Automation
• Loss of Control Training
• Airport Surface Operations Training
• Line Operations Safety Audit (LOSA)
• Training for In-flight fires
Why Monitoring?
October 25, 2002 Eveleth, Minnesota
 According to NTSB
• “during the later stages of the
approach, the flight crew failed to
monitor the airplane’s airspeed and
allowed it to decrease to a dangerously
low level (as low as about 50 knots
below the company’s recommended
approach airspeed) and to remain
below the recommended approach
airspeed for about 50 seconds.”
Source: NTSB/AAR-03/03.
• “If the First Officer had monitored the
approach on the instruments...he would
have been better able to notice and
immediately call the Captain’s attention
to the altitude deviation below the
minimum descent altitude.”

Source: NTSB/AAR-96/05.
 Chief Test Pilot
• June 30, 1994
• Wide-body aircraft
• Preparation for certification of autopilot for Cat 3
landing weather minima
• Shortly after takeoff, autopilot was engaged, followed
by simulated engine failure
• Pilots became involved with executing test plan,
turning off hydraulic system
• Speed decay was not noticed, aircraft slowed below
Vmc
• 7 fatalities
 Why improve monitoring?
- Accident Data -
• Inadequate crew monitoring or challenging was a factor
in 84% of 37 crew-caused air carrier accidents reviewed
in a NTSB safety study.
– 76% of the monitoring/challenging errors involved failure to
catch something that was causal to the accident
– 17% of the monitoring/challenging errors were failure to catch
something that contributed to the accident’s cause
• Poor monitoring was a factor in 63% of the ALA
accidents reviewed by the FSF ALAR working groups.
• 50% of the CFIT accidents reviewed by ICAO to support
the FSF CFIT efforts involved poor monitoring.
Why is monitoring important ?
- LOSA Data -
• Roughly 64% of “unintentional errors” in the
University of Texas LOSA archive were
undetected by flight crew.
• In a recent LOSA, 19% of errors could have
been eliminated by more effective crew
monitoring and cross-checking.
• In that same LOSA, 69% of “undesired states”
could have been eliminated by more effective
monitoring.
Concerns about Deficiencies
in Monitoring
“The HF Team has concerns that incidents
demonstrating deficiencies in flightcrew
monitoring and awareness of autoflight
system modes, airplane energy state, terrain
proximity, and airplane systems’ status are
occurring to an unacceptable extent.”

– FAA Human Factors Team Report on the Interfaces Between

Flightcrews and Modern Flight Deck Systems, page 44.
Role of Design in Improving
Monitoring
“Design and training aspects that influence the ability
of flightcrews to maintain situation awareness must
also be addressed.”

“Mode changes can easily be missed, even when

additional cues are provided (e.g., drawing a box
around the new mode and/or using a flashing display
for a few seconds), unless a flightcrew member is
looking at the display when the change occurs. With
the autopilot on, flightcrew members are often not
looking at the PFD when a mode transition occurs.”

- FAA Human Factors Team Report on the Interfaces Between Flightcrews

and Modern Flight Deck Systems, page 44.
• “We missed the crossing altitude by 1000
feet. The captain was...busy trying to
program the FMC. Being new in an
automated cockpit, I find that pilots are
spending too much time playing with the
computer at critical times rather than flying
the aircraft.”

– NASA ASRS database

 US Airways decided to
actively work to improve
pilot monitoring
This is a collaborative effort
between US Airways, ALPA and
NASA Ames Research Center
Underlying factors associated
with poor monitoring

1. Until now, the industry has not made

monitoring a primary task.
– When listing PNF duties, we often list duties
such as handling radio communications,
keeping flight logs and operating gear and
flaps.
– Monitoring is not one of the duties primarily
listed, but rather it seems to be treated as a
secondary task, or not addressed at all.
Underlying factors associated
with poor monitoring

2. Effective monitoring is not easy and

intuitive
– requires a skill and discipline
Underlying factors associated
with poor monitoring
3. There is somewhat of a monitoring
paradox that works against effective
monitoring.
– Serious errors do not occur frequently which
can lead to boredom and complacency

“A low-probability, high-criticality error

is exactly the one that must be
caught and corrected.”
Underlying factors associated
with poor monitoring

4. Although traditional CRM courses have

generally improved the ability of
crewmembers to challenge others
when a situation appears unsafe or
unwise…
• many of these courses provide little
or no explicit guidance on how to
improve monitoring.
Underlying factors associated
with poor monitoring

5. We seem satisfied that we cannot

improve monitoring, and simply
explain it as, “Humans just are not
good monitors.”
– While it may be true that humans are not naturally
good monitors, crew monitoring performance can be
significantly improved through policy changes and
crewmember training.
NASA ASRS Monitoring Study

• This study was a good

resource for helping us get
stared and providing data
What ASRS Data • The objectives of the study
Tell About Inadequate were to identify factors that
Flight Crew Monitoring contribute to monitoring
errors, and
• Offer operationally- oriented
recommendations to
– increase awareness of this
subject
– improve crew monitoring
 Number of monitoring errors

0
10
20
30
40
50
60

taxi-out

takeoff

climb

cruise

crs/des trans.

descent

holding

approach

landing

taxi-in
Flight Phase where
ASRS Monitoring Study:

Initiation
Monitoring Errors Occurred
ASRS Study: Number of tasks crew
was doing when error occurred*
Reports
*As reported
89%
140
in 200 reviewed
120 ASRS reports
100
80
40%
60
40
20
0
0 1 2 or more
Number of tasks crew was doing shortly before
or during the initiation of the monitoring error.
 ASRS study significant
findings
• 76 percent of monitoring errors
occurred when aircraft was climbing,
descending or on approach (“vertical
flight phase”)

• 30 percent of the reports indicated that

pilots were programming the FMS
shortly before or during the monitoring
error
 Which Mode Causes the
Most Confusion or Errors ?
• MIT analysis of 184 mode awareness
incidents found

Confusion or errors

– Vertical Nav 74 percent

– Lateral Nav 26 percent
US Airways Approach to
Improving Monitoring

Developing well thought-out SOPs

Training monitoring skills
Practicing those skills
US Airways Approach to
Improving Monitoring

Developing well thought-out SOPs

Training monitoring skills
Practicing those skills
 Developing SOPs

AC 120-71A “Standard Operating Procedures”

• Revised in February 2003, this AC

contains template SOPs that can be
adopted by operators to improve
monitoring.

www.aviatrends.com
 Developing SOPs

• Change title of “Pilot-Not-Flying” (PNF)

to “Pilot Monitoring” (PM)

– Describes what the pilot should be doing

(monitoring) versus what he/she is not doing
(not flying)
 Developing SOPs

• During high workload, FMS inputs will

be made by PM, upon the request of
PF.
High workload examples
– below 10,000 feet
– within 1000 feet of level off or Transition
Altitude.
 Developing SOPs

• Perform non-essential duties/activities during lowest

workload periods (e.g., cruise altitude or level flight)
• When able, brief anticipated approach prior to top-of-
descent
• PF will brief PM where or when delayed climb or
descent will begin
• During the last 1000 feet of altitude change, both pilots
will focus on making sure the aircraft levels at the
assigned altitude
• US Airways eliminated “10,000 ft announcements” and
specified that the pre-arrival announcement be
conducted just after leaving cruise altitude.
US Airways Approach to
Improving Monitoring

Developing well thought-out SOPs

Training monitoring skills
Practicing those skills
Training monitoring skills

“First, we must change our approach to

monitoring. Instructors must [teach and]
insist that the non-flying crewmember
monitors the flier effectively.”
“Good monitoring skills are not inherent in
pilots as they progress in their careers.
Therefore, effective monitoring techniques
must be trained and rewarded.”

- Captain Frank J. Tullo

“Aviation Week and Space Technology,” May 21, 2001
Training monitoring skills
• NTSB safety study states that simulator
training provides a good opportunity to
teach and practice monitoring and
crosschecking.
– NTSB Safety Study of Crew-caused Accidents
 Training monitoring skills
• Starting from day 1 of training, ensure all monitoring/
crosschecking SOPs are followed.
• Discuss how barriers are cut in half with one pilot out of the
loop.
• Train workload management so at least one pilot is always
monitoring during low workload and both pilots are
monitoring as much as possible during high workload.
• Acknowledge good monitoring.
– Introduction of occasional subtle failures in simulator training,
such as failure of automation to level-off at proper altitude
US Airways Approach to
Improving Monitoring

Developing well thought-out SOPs

Training monitoring skills
Practicing those skills
Practicing monitoring skills
• Know and comply with SOPs
• Pilots must “actively monitor” the aircraft.
• This means that they must mentally fly the
aircraft, even when the autopilot or other
pilot is flying.
– Monitor the flight instruments just as you would when
hand flying.
– If the aircraft (or other pilot) is not doing what it is
supposed to do, actions should be taken to rectify the
situation.
 Practicing monitoring skills
• In approximately one-third of the cases studied
by researchers, pilots “failed to monitor errors,
often because they had planned their own
workload poorly and were doing something else
at a critical time.”
– Jentsch, Martin, Bowers (1997)
• Threat and Error Management with a focus on
monitoring and cross-checking is a good way to
teach pilots better workload prioritization skills.
Practicing monitoring skills
• Pilots should recognize those flight phases
where poor monitoring can be most
problematic.
• Strategically plan workload to maximize
monitoring during those areas of vulnerability
(AOV)
– Examples of non-monitoring tasks that should be
conducted during lower AOV include stowing charts,
programming the FMS, getting ATIS, accomplishing
approach briefing, PA announcements, non-essential
conversation, etc.
 Practicing monitoring skills
Areas of Vulnerability
Descent,
Within 1000 ft
of level-off Approach
and Landing
Transition alt Cruise-Descent
Transition, or
anytime you are
10,000 ft
anticipating a clearance

Taxi-in
Taxi-out
Practicing monitoring skills
• By pre-briefing the approach in low workload
periods, greater attention can be devoted to
monitoring/cross-checking during descent.
• In fact, US Airways LOSA data showed that
crews who briefed the approach after Top-Of-
Descent (TOD) committed 1.6 times more
errors during the descent/ approach/land
flight phase than crews who briefed prior to
TOD.
Practicing monitoring skills
• One way of assessing your current
monitoring ability is to ask: “How often
do I miss making the 1,000’ to level-off
altitude callout?”
– When this callout is missed, chances are that
you are not actively monitoring the aircraft.
Paradigm shift
• It must become accepted
that monitoring is a “core
skill,” just as it is currently
accepted that a good pilot
must posses good “stick
and rudder” and effective
communicational skills.
 Safety Order of Precedence
1. Design for Minimum Risk (engineering
solution)
– Hazard is corrected and eliminated
2. Control/Guard Solution
– Guards put up to decrease exposure
3. Personnel Warning System
– Warn personnel if you can’t eliminate or control the hazard
4. Develop Procedures and Training
“… these concerns extend beyond
deficiencies in crew’s monitoring
techniques, [training aspects] and
adherence to procedures. Design …
aspects that influence the ability of
flightcrews to maintain situation
awareness must also be addressed.”
New 25.1329 Rule (Draft)
25.1329 (i)
(i) The FGS functions, controls, indications, and
alerts must be designed to minimize flight crew errors
and confusion concerning the behavior and operation of
the FGS. Means must be provided to indicate the
current mode of operation, including any armed
modes, transitions, and reversions. Selector switch
position is not an acceptable means of indication. The
controls and indications must be grouped and presented
in a logical and consistent manner. The indications must
be visible to each pilot under all expected lighting
conditions. Crew Awareness