What’s New in Human Factors?

1
Whats New in Human Factors?

• Kim Cardosi
• Cardosi_at_volpe.dot.gov
• September 12, 2006

2
Intro and Disclaimer

• DOT/research not FAA
• PhD in Experimental Psychology (not clinical or
  MD).
• Views are my own and not those of the FAA or DOT.
• Nothing in this presentation should be
  interpreted as a change in policy or procedure.

3
Todays Topics

• Latest research findings
• Proposed changes to how we measure safety
  performance in the NAS.
• Calling for a (safety) cultural revolution.

4
Research Findings of Interest

• 10-year update on TRACON controller-pilot voice
  communications showed
• Comparable rate of controller-pilot
  transmissions/clearances per minute.
• Comparable rate (1) of readback errors.
• Substantial increase in full readbacks with call
  signs.(60 to 83).
• Readback errors and requests for repeats increase
  with information complexity. 3
• Transmissions that contained more than 3 pieces
  of information - accounted for 26 of the
  readbacks and 51 of the readback errors.

5
Pilot Behavior

• Most common Pilot error that results in in
  runway incursions is STILL
• Pilot reads back instruction correctly and then
  does something else usually cross the hold
  short lines.
• Most common result is a go-around.

6
Lights on for Takeoff

• AC 120-74A was issued instructing Part 121 pilots
  to reserve the use of the landing light for
  takeoff.
• Since then the most serious (A and B) incursions
  involving crossing/ potential crossings in front
  of a takeoff has decreased by 21.
• 47 (2000-2002) to 26 (2004), 23 (2005).
• Exclude those incursions caused by vehicle
  drivers and 23 drops to 15 (in 2005).
• Getting real and going global.

7
Most populated controller error
categoriesGotchas

• Tower
• Forgot (runway closed, aircraft on approach,
  aircraft on the runway, approved a crossing)
• Inadequate coordination
• ARTCC/TRACON
• Overlooked traffic
• Readback/hearback
• Inadequate coordination
• OEs incurred by DEV result in aircraft getting
  closer together.

8
Weakest Links

• Prospective memory trying to remember to do
  something in the future is the most fragile part
  of short-term memory.
• Time (decay)
• Interference - Distractions
• Habit (human auto pilot)

9
Holding on the Runwayfor imminent take off 2
minutes - too long
10
Examples of Mitigation Strategies

• Standardized procedures
• e.g., checklists
• Memory aids
• CRM Crew Resource Management

11
ATC Crew Resource Management (CRM)

• Improving Teamwork giving the traffic youd
  like to get
• Improving Individual Performance recognizing
  danger zones
• Error Management
• Workforce ideas are recorded and put into a
  facility feedback document, to be used for
  follow-up and action plans.

12
Safety Culture

• Just culture - honest mistakes are treated as
  such.
• Reporting of adverse events is encouraged.
• Investigation of adverse events looks at the
  PROCESS, not just the person.
• Accountability at all levels.
• Track progress.

13
Goals of Revised Safety Measures

• To develop measures that are operationally useful
  and scientifically defensible.
• Provide specific measures of system performance.
• Establish baselines to assess the effects of
  changes (e.g., equipment, airspace, procedures,
  standards, facilities)
• Reliable - remove all subjectivity.
• To separate process from outcome
• SYSTEM outcome

14
A Tale of Two As

• 3/28/05 OE Mobile ATCT AC1, Beech BE9L was
  landing at JKA and AC2, Beech BE58, departing JKA
  were on opposite direction flight paths. The CPC
  had issued AC1 a descent to 3000 feet and had
  climbed AC2. The OS sitting at the adjacent
  position saw the conflict and asked the CPC to
  make sure his plan was working. AC1 had asked to
  go to the ILS FAF which the CPC approved thinking
  it would take the aircraft out of the way. The
  FAF though was not where he thought it was. The
  CPC made another transmission and when he
  instructed AC2 to turn the pilot did not answer.
  He then instructed AC1 to turn, which did happen.
  The targets merged with AC1 responding to a TCAS
  RA climb, which provided 300 feet. Closest
  proximity 0 miles and 300 feet. High Severity
  (Category A). MOB-T-05-E-001
•  

15
A Tale of Two As (Contd)

• 5/26/05 New York ARTCC RA was in place giving
  and taking handoffs. DEVs 1st session at a radar
  sector. FL180 was unusable so numerous aircraft
  were at FL190 adding to the complexity. OJTI
  took the frequency. AC1CHQ3063, E145 was S-bound
  for PHL at FL190 and AC2CAA795, CRJ was NE-bound
  for ABE at FL190. Both were on headings to miss
  other traffic as they converged over Harrisburg.
  Conflict Alert went off and CAA795 responded to
  TCAS RA by climbing to FL199. Closest proximity
  300 feet and 3.25 miles. High Severity (Category
  A). ZNY-C-05-E-020

16
Severity IndexA Tale of Two Bs

• 4/4/06 Fort Worth ARTCC Aircraft 1 (B752) was
  east bound at FL370, and Aircraft 2 (C56X) was
  offset opposite direction at FL380. CPC failed to
  recognize AC1 as traffic and issued AC2 a descent
  for landing FTW. CPC turned AC2 30 degrees to the
  right but it was not soon enough to mitigate the
  loss. Closest Proximity same altitude and 4.58
  miles,

17
Severity IndexA Tale of Two Bs

• 5/26/06 Falmouth MA TRACON AC1 departed MVY
  climbing to 10,000 initially heading southbound
  and then turned northeast bound. AC2 departed
  ACK northwest bound climbing to 10,000. The CPC
  issued an expedited climb to 10,000 to AC1 for
  other traffic, handed the aircraft off, and
  switched it to ZBW. Both aircraft were on a
  converging course towards the MVY VOR. Conflict
  Alert activated and AC2 responded to a TCAS RA.
  AC1 advised ZBW they would file a NMAC. Closest
  proximity 500 feet vertical, .36 miles lateral.
  Moderate Severity, Category B.

18
(No Transcript)
19
(No Transcript)
20
Process and Outcome

• By separating process from outcome, we can relate
  factors that contributed to the losses of
  standard separation with the severity of the
  outcome.
• Outcome the magnitude of the loss of separation
• Process examines the situational factors and
  human errors that resulted in the loss of
  standard separation

21
Categories of Factors

• Organizational (staffing, supervisory, local
  procedures)
• Scenario (conflict geometry, traffic
  level/complexity)
• System architecture (communications, situation
  display)
• Controller Errors (e.g., readback/hearback.
  overlooked aircraft, coordination error, issued
  an instruction other than intended)
• Conflict detection and resolution (who, when,
  how)

22
Human Error is Inevitable

• Even the most experienced, attentive,
  well-trained, well-rested, most conscientious
  professional can make a mistake.
• It isnt reasonable to ask that we achieve
  perfection. What is reasonable is that we never
  cease to aim for it.
• Atul Gawande Complications