1990s
2020s
1960s
DESIGN OF A SINGLE PILOT
COCKPIT FOR AIRLINE
OPERATIONS
Jonathan Graham, Chris Hopkins, Andrew Loeber, Soham Trivedi
Overview





Problem: Poor financial performance in commercial aviation and predicted
pilot shortage
Need: System needed to reduce airline costs and hedge potential pilot
shortages
How: Single Pilot Cockpit system potentially reduces pilot labor need and
airline labor cost
Our Job: Analyze design alternative’s ability to meet system need within
project scope and stakeholder win-win
Outcomes: Recommend systems based on the derived feasibility of
designing a Single Pilot Cockpit
2
Agenda








Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
3
Scope

Large Commercial Air Transportation




US Airspace




Passenger and Cargo Carriers
Carriers With Operating Revenue >$20 Million
Domestic Operations
National Airspace System (NAS)
ATC
FAA Regulatory Body
Financial data adjusted for inflation to 2012
dollars
4
[1]
Profit/Loss for Major Air Carriers
20.00
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
-10.00
1991
0.00
1990
Dollars (Billions)
10.00
Profit/Loss
Net Income
-20.00
Bankruptcies Filed
-30.00
-40.00



Year
Large air carriers as a whole have had volatile financial
performance
The year 2000 was the tipping point
30% of US Airlines filed for Chapter 11 between 2000-2010
[1] BTS Schedule P1.2
*Values are inflation adjusted to 2012 based on consumer price index
5
[2]
Projected Total Operating Expense for Major Air Carriers
Operating Expense (Billions)
200.00
180.00
160.00
Total
Operating
Expense
140.00
120.00
100.00
Projected
Total
Operating
Expense
80.00
60.00
40.00
20.00
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
0.00
Year


Projected growth in operating expense
Reducing operating expense relative to operating
revenue is good for financial stability
[2] BTS Schedule P5.2
*Values are inflation adjusted to 2012 based on consumer price index
6
% Operations Expense for Major Air Carriers
[3]
60.00
% Operations Cost
50.00
40.00
% Pilot
30.00
% Fuel
20.00
% Dir Maint
10.00
% Air Ops
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
0.00
Year

Large percentage of operating expense is composed of
fuel and pilot labor costs
Fuel costs are a variable cost
 Pilot labor costs are easier to control

[3] BTS Schedule P5.2
*Values are inflation adjusted to 2012 based on consumer price index
7
[4]
Projected Pilots
80
Pilots (Thousands)
70
60
50
40
Total Pilots
30
Projected
Pilot Demand
20
Projected
Supply
10
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
0
Year





The number of pilots has been relatively static for the last decade
Pilot labor growth has been much greater in the past
Flight hour requirements and decreased retirement age impacting future pilot supply
6% Growth in commercial pilot labor projected from 2012-2022 using FAA forecasts for 2013-2033 [9]
~4,500 Air Transport Pilot licenses per year from 2004-2012 [5]
[4] BTS Schedule P10
[5] WSJ Airlines Face Acute Shortage of Pilots, Carey et al.
*8.94% attrition rate and fixed licensure rate used to plot shortage curve in red [11]
8
Radio
Flight
Navigator
Co-Pilot?
Operator
Engineer
Is a Single Pilot Cockpit the Next Step?
9
The Flight Deck

Two pilots operate the aircraft

Pilot Flying



Pilot Not Flying





Flies the Aircraft
Confirms Callouts
Inspects/Manipulates Instruments
Performs Callouts
Interacts with ATC
Flies on Behalf of Pilot
Both captain and co-pilot can take on each role during a flight
10
Cockpit Avionics
Mode Control Panel
Center Instrument Panel
1st Officer’s Instrument Panel
Avionics
Downlink
Control Display Unit
Overhead Panel
11
Flight Procedures


Procedures are followed to operate an aircraft
Flight Crew Operating Manual (FCOM) BAE RJ100



Identifies responsible entities for procedural tasks



Describes flight procedures
Official FAA approved document
Pilot Flying (PF)
Pilot Not Flying (PNF)
Why do we care?

Used in analysis to show how workload is affected
when alternatives change the operating procedures
12
Procedure Decomposition



A procedure is decomposed by tasks and
physical/cognitive actions that are required to
execute a task
Procedure
FCOM Codifies Procedures

63 Procedures

613 Total Tasks

737 Total Actions
Procedures are decomposed to identify potential
reallocation of actions to a design alternative
Tasks
Actions
13
Example Procedure
14
Action Frequencies
180
160
140
120
100
PF
80
PNF
60
B/P
40
20
0
Physical
Instrument
Manipulation
Verbal Cockpit
Callout
Physical Flight
Computer
Interaction
Auditory
Reception
Memory Action
Visual Instrument
Inspection
Visual
Environment
Inspection
Verbal External
Communication
% Actions by Pilot
B/P
5%
PF
40%
PNF
55%
15
15
Agenda








Background & Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
16
Stakeholder
Group
Regulatory Agencies
Primary
Objectives
•
Maximize:
• Flight safety
• Consumer protection
A SPC would inherently introduce new risks and
decrease overall flight safety, leading regulatory
agencies to withhold their approval
•
Maintain:
• Job Stability
• Wage Stability
• Safety level
• Workload
View SPC as a major potential threat to job stability,
leading to a high risk of pushback
•
Minimize:
• Travel time
• Flight risk
• Ticket expenses
May have reservations about flying in a plane with
only one pilot, leading them to avoid flying with an
airline that uses SPC
•
Maintain:
• Consistent revenues
• Customer base
• Market predictability
• Low risk profile
Want to increase profitability through sales/service,
but don’t want to increase expenses commit to long
term investments without noticeable return
1717
(FAA, DoT)
Aviation Workforce
(Pilots, Pilots’ Unions, ATC, ATC Unions)
Customer Base
Aviation Industry
(Air Carriers, Management,
Manufactures, Insurance, & Airports)
Tension with SPC
Stakeholder Interactions
Support SPC
Reserved about SPC
Oppose SPC
Destabilizing
Stabilizing
Neutral
18
18
Agenda








Background & Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
19
Problem Statement

Rising operating expense contributes to financial
instability in commercial aviation
 Profitability

is difficult to achieve
Pilot labor shortage predicted
20
GAP Analysis
Operating Revenue to Operating Expense Ratio
*BTS Schedule P5.2
1.4
1.2
1
Ratio
0.8
0.6
Ratio
0.4
Target
Ratio
0.2
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Year
*Values are inflation adjusted to 2012 based on consumer price index
21
Need Statement

Airlines will continue to have volatile financial
performance if operating costs continue to grow
 Variable
costs like fuel driving operating expense
 Labor expense can be controlled more effectively

A single pilot cockpit is needed to decrease labor
expense and mitigate the effects of a pilot
shortage
22
Organization
Tension to be
Mitigated
Benefit to Slow Phase-In
Fear of elevated risk
caused by removing a
pilot
Allows regulatory agencies to observe the effects of implementing a
SPC and collect reliability data without the worry of deploying an
uncertain system and dealing with damage control.
Fear of labor downsizing
The resultant decline in pilot labor demand can be spread out over
several decades, meaning that job stability can remain relatively
stable, and pilots can adapt to using a new system. A SPC system can
also potentially reduce a pilot’s workload.
Customer Base
Fear of boarding a
plane being flown by a
single pilot
Fliers with concerns about the safety of a SPC will be allowed more
time to acclimate to the new technology. Also, the majority viewpoint
will shift due to changing generational attitudes regarding automation
in general.
Aviation Industry
Fear of costs/changes
needed to adapt to new
system
Airports and aircraft manufacturers will be given additional time to
adapt their operations, products, and business plans to the current
phase of SPC deployment, keeping them from wasting resources on
developing unutilized solutions.
23
Regulatory
Agencies
Aviation
Workforce
Win-Win Analysis


Majority of stakeholders involved have serious conflicts with moving to
a single-pilot system
However, these issues can be mitigated by extending the phase-in
process



Helps to minimize safety concerns brought on by major systemic change
Decreases financial risk
Gradual labor downsizing
Today
5 Years
10 Years
15 Years
Two Pilot
Cockpit
Two Pilot
Cockpit with
Alternative
Evaluation of
Alternative
Single Pilot
Cockpit
24
Agenda








Background & Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
25
Requirements
Requirement ID
Description
M.1
The single pilot cockpit system shall reduce or maintain the baseline
pilot flying procedure working time of 20.87+-3.7s
M.2
The single pilot cockpit system shall meet ARP4761 Level B assurance
of 1 failure per million flight hours.
M.3
The single pilot cockpit system shall decrease yearly pilot labor
operating expense.
M.4
The single pilot cockpit system shall have a total aircraft lifecycle cost
no greater than $153.9 million dollars.
26
Agenda








Background & Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
27
Physical Process Diagram
ATC
Focus of
Analyses
Flight Goal
Operating
Procedures
Pilots
System
Selected Procedures
Cockpit
Aircraft Control
Aircraft
Flight
Outcome
Avionics
Aircraft Dynamics
28
Design Alternatives
1.
2.
3.
Two Pilot Cockpit (No Change)
Single Pilot with No Support
Single Pilot with Onboard Support System
29
Agenda








Background & Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
30
31
Procedure Simulation
Reliability Modeling
Business Case
Assumptions







Human Processor Model approximate operator actions in cockpit
under normal flight conditions
Events are independent of each other
Normal flight conditions and expert skill level
Assume alternatives follow contemporary avionics costs
RJ100 FCOM is representative of similar operating manuals
compiled by commercial airlines
Assume procedures in operating manual are complete
representation of flight
Additional company specific pilot tasks not included
[12] Liu, Feyen, Tsimhoni: Queuing Network Model Human Processor
32
Procedure Simulation
Reliability Modeling
Business Case
Input Methodology


Input data derived from RJ100 Flight Crew Operating
Manual (FCOM)
Convert FCOM and classify procedures based on ontology



Procedures are changed for each alternative
One procedure model for each alternative
Each alternative’s procedural model is translated into an
XML representation

XML is parsed into simulation
33
Procedure Simulation
Reliability Modeling
Business Case
Simulation Construction

Java program



Input procedural model for each alternative
Output Alternative Processing Time (APT)
Procedure Model
Simulation
Processing Time
Simulation based on Model Human Processor




Network of perceptual, cognitive, and motor
sub-networks
Each process associated with capacity and
process time
Model combines concepts from ACT-R, GOMS,
and Fitt’s Law
Additional block added for alternative
processing time where applicable
[12]
[12] Liu & Wu: Modeling Psychological Refractory Period (PRP) and Practice Effect on PRP with Queuing
Networks and Reinforcement Learning Algorithms
34
34
Inside the Simulation
[12] Liu & Wu: Modeling Psychological Refractory Period (PRP) and Practice Effect on PRP with Queuing Networks and Reinforcement Learning Algorithms
50ms
.
.
.
.
Procedure j
Task i
Action j
Eyes
N(263ms,11ms)
Exp(50ms)
2ms
41ms
VSen
CE
PM
Hands
.
.
.
Action n
Task n
Procedure n
70ms
M1
10ms
50ms
Ears
ASen
Perception
180ms
1ms
10ms
SMA
BG
Mouth
Cognition
Motor
35
Procedure Simulation
Reliability Modeling
Business Case
Equations
Simulation
Perception
Cognition
Motor
Eyes
VSen
Ears
ASen
CE
SMA
M1
PM
50ms
N(263ms,11ms)
10ms
50ms
Exp(70ms)
180ms
70ms
A+Bexp(-αN)
Replications
BG
Hands
kLog(D/S+0.5)
Mouth
10ms
Analysis
𝐻0 : 𝐴𝑃𝑇𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 = 𝐴𝑃𝑇𝐶𝑜𝑛𝑡𝑟𝑜𝑙
𝑊𝑜𝑟𝑘𝑙𝑜𝑎𝑑 =
𝑡𝑖𝑚𝑒 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒𝑖..𝑗
=
𝑡𝑖𝑚𝑒 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒 𝑇𝑤𝑜 𝑃𝑖𝑙𝑜𝑡 𝑃𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒𝑘..𝑙
𝐻1 : 𝐴𝑃𝑇𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 > 𝐴𝑃𝑇𝐶𝑜𝑛𝑡𝑟𝑜𝑙
𝑟𝑒𝑗𝑒𝑐𝑡 𝑖𝑓𝑓 𝑝 < 0.05
[12] Liu, Feyen, Tsimhoni: Queuing Network Model Human Processor
36
36
37
Procedure Simulation
Reliability Modeling
Business Case
Design of Experiment
Input
Output
Alternative
Procedures
& Tasks
Additional Processing Node
Single Pilot No Support
P1…Pn
T1m…Tnm
None
Two Pilot
P1…Pn
T1o…Tno
None
Single Pilot With Onboard Support
System
P1…Pn
T1p…Tnp
Exp(1/X)
Processing Time
Workload
37
Procedure Simulation
Reliability Modeling
Business Case
Concept

Simple reliability block diagram used to establish comparative
reliability between cockpit elements and major accidents





Major accidents as defined by NTSB
Major accidents per million flight hours ~ MTBF
Assume that cockpit reliability is critical to aircraft and can be isolated
from other subsystems
Assume there is an emergency land capability in the event of pilot
incapacitation
Don’t know the reliability of the system, but assists in reliability
requirements development
[15] NTSB Accident Data: http://www.ntsb.gov/data/aviation_stats.html
38
Procedure Simulation
Reliability Modeling
Business Case
39
RBD and Formulas
Pilot
CoPilot
Pilot
Two Pilot
Aircraft
Aircraft
Single Pilot
Onboard
Support
Pilot
Aircraft
Single Pilot with Onboard
39
Support
Procedure Simulation
Reliability Modeling
Business Case
Design of Experiment
Input
Output
Alternative
Cockpit Failure Rate
(per flight hour)
Aircraft Failure Rate
(per flight hour)
Target System
Reliability (@1
Million Flight Hours)
Required Reliability
Single Pilot No Support
λc = (100k-1,10m-1)
λA = (100k-1,10m-1)
0.8674
Rs = RARP1
Two Pilot
λc = (100k-1,10m-1)
λA = (100k-1,10m-1)
0.8674
Rs = [1-(1-RP1)(1-RP2 )]RA
Single Pilot With Onboard
Support System
λc = (100k-1,10m-1)
λA = (100k-1,10m-1)
0.8674
Rs = [1-(1-Ralt)(1-RP)]RA
[15] NTSB Accident Data: http://www.ntsb.gov/data/aviation_stats.html
40
Procedure Simulation
Reliability Modeling
Business Case
Concept


Model an aircraft lifecycle cost (25 years) for each
alternative and calculate net savings (if any)
Limited aircraft costs to:
 Maintenance
data collected for Boeing 737s
 Average pilot labor costs used with an escalation rate
 Fuel too variable for our analysis
 Salvage value of aircraft considered at end of life
41
Procedure Simulation
Reliability Modeling
Business Case
Formulas
Cost Formulas
42
42
Procedure Simulation
Reliability Modeling
Business Case
Design of Experiment
Input
Alternative
Single Pilot
No Support
Alternative Cost
Pilot Cost
Output
Maintenance
Cost
Residual Cost
Interest
CM = $9.127M
CR =$8M-$9M
d=1%-10%
CP = $141,927
CA = $0
E=0.05%-5%
CP = $269,126
Two Pilot
CA = $0
CM = $9.127M
CR = $8M-$9M
d=1%-10%
CM = $9.127M
CR = $8M-$9M
d=1%-10%
Total
Lifecycle
Cost
Net Savings
E=0.05%-5%
Single Pilot
With
Onboard
Support
System
CP = $141,927
CA = ($100K,$4.38M)
E=0.05%-5%
43
Value Hierarchy
Single Pilot
Utility
Alternative
Processing
Time
In Scope
Out of Scope
Workload
Risk
Technology
Readiness
Level
Reliability
Training
Liability
Maintenance
Usability
44
Agenda








Background & Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
45
Procedure Simulation

Seconds
Sample Average Procedure Times
35.00
30.00
25.00
20.00
15.00
10.00
5.00
0.00

Two Pilot
Single Pilot
Pilot Support
1
2
3
4
5
6
7
Procedures
8
9
10

Predicted reduced processing
times due to reduction in
procedures
Both single pilot cockpit
designs on average have
procedure times smaller than
the two pilot cockpit
p<0.0001
Time pilot spends performing
tasks for operating
procedures is reduced
46
Procedure Simulation
Frequency (Thousands)
Procedure Processing Time
Distributions
25
20
15
Single Pilot
10
Two Pilot
5
Pilot Support
0
0 15 30 45 60 75 90 105 120 135 150
Time (s)
Alternative
(n=1000 trials)
Mean
(s)
Median
(s)
Standard
Deviation
(s)
Two Pilot (No Change)
20.868
14.193
20.926
Single Pilot
12.703
9.386
13.770
Single Pilot with Onboard
Support
16.909
11.894
18.637
47
Procedure Simulation
35

Single Pilot Cockpit Workload
30

Workload
25
Single
Pilot
Workload
Pilot
Support
Workload
20
15
10
5
0
Procedures

Single pilot had smallest
processing time p<0.0001
The single pilot incurred a
significant increase in
workload: 4.17+/-1.77
Single pilot with onboard
support had a reduction in
workload: 0.75 +/-0.13
48
Reliability Analysis

Cockpit Reliability Requirement
Single pilot would not be achievable
to maintain required reliability
1
0.99
0.99-1
0.98
0.98-0.99
0.97
0.97-0.98
0.96
10
7
0.95
0.94
1
4
2
3
4
5
6
7
8
9
1
0.96-0.97
MTBF 2
Reliability (1 Million Hrs)

0.95-0.96
0.94-0.95

~1 failure every23 million flight
hours
Single pilot with onboard support
would be feasible as long as an
emergency auto-landing feature
could be integrated into the system

Avionics would have to be certified
above 1 in 1 million flight hour failure
10
MTBF 1
49
Business Case

Single Pilot Savings
$4.00
$3.00
$2.00
$1.00
$0.00
0.03
0.005
Interest Rate
Pilot Cost Escelation Rate
NPV (Millions)
$5.00
$4.00-$5.00

$3.00-$4.00
$2.00-$3.00
$1.00-$2.00
$0.00-$1.00

Single pilot would support
up to a $4.38M peraircraft savings
Savings could be
allocated to procedure
support system acquisition
or other operating costs
Single pilot cockpit
decreases aircraft
operating costs
50
Utility
Utility vs. Cost
1.20
Utility
1.00
Two Pilot Reliability Max
Single Pilot Reliability Max
0.80
Procedure Support Reliability
Max
Two Pilot Workload Max
0.60
Single Pilot Workload Max
0.40
Procedure Support Workload
Max
Two Pilot Procedure Time Max
0.20
Single Pilot Procedure Time Max
0.00
$149.00 $149.50 $150.00 $150.50 $151.00 $151.50 $152.00 $152.50 $153.00 $153.50 $154.00 $154.50
Cost (Millions)
Procedure Support Procedure
Time Max
51
Agenda








Background & Context
Stakeholder Analysis
Problem & Need
Requirements
Design Alternatives
Simulation & Methodology
Results
Recommendation
52
Requirements
Alternatives Satisfying Requirement
Requirement ID
Description
Two Pilot
Single Pilot No
Support
Single Pilot
w/ Onboard
Support
System
M.1
The single pilot cockpit system shall reduce or
maintain the baseline pilot flying procedure working
time of 20.87+-3.7s



M.2
The single pilot cockpit system shall meet ARP4761
Level B assurance of 1 failure per million flight hours.
n/a
X

M.3
The single pilot cockpit system shall decrease yearly
pilot labor operating expense.
X


M.4
The single pilot cockpit system shall have a total
aircraft lifecycle cost no greater than $153.9 million
dollars.
X


53
Recommendation

Utility Ranking:
1.
2.
3.
Two Pilot Cockpit
Single Pilot with Onboard Support
Single Pilot
A single pilot has cost reductions but workload and safety bounds overall
utility


Safety mechanism for pilot would give alternative increased utility (and cost)
A single pilot with onboard support has marginally equivalent utility to
the two pilot cockpit but has cost savings


Larger cost savings from unexplored areas could lower costs making utility
ranking highest
54
Recommendation


Recommend keeping two pilot cockpit and evolve
alternative system per the win-win scenario
Procedure Support system to be phased in and
evaluated for the eventual change to the single pilot
cockpit pending further future analysis
Today
5 Years
Two Pilot
Cockpit
Two Pilot
Cockpit with
Alternative
10 Years
Evaluation of
Alternative
15 Years
Single Pilot
Cockpit
55
Further Research

Extend procedure simulation to a live pilot simulator to
validate and/or recalibrate models
Tie in ATC procedures to test their workloads
 Incorporate dynamics feedback loop



Conduct focus groups with stakeholders to validate winwin scenario
Begin development of Single Pilot with Onboard
Support requirements baseline
56
Questions?
57
Sources
[1] BTS Schedule P1.2 http://www.transtats.bts.gov/Fields.asp?Table_ID=295
[2] BTS Schedule P5.2 http://www.transtats.bts.gov/Fields.asp?Table_ID=297
[3] BTS Schedule P5.2 http://www.transtats.bts.gov/Fields.asp?Table_ID=297
[4] BTS Schedule P10 http://www.transtats.bts.gov/Fields.asp?Table_ID=302
[5] WSJ Airlines Face Acute Shortage of Pilots
http://online.wsj.com/news/articles/SB10001424052970203937004578079391643223634#ar
ticleTabs%3Darticle
[6] BTS Schedule P10 http://www.transtats.bts.gov/Fields.asp?Table_ID=302
[7] BTS Schedule P5.2 http://www.transtats.bts.gov/Fields.asp?Table_ID=297
[8] BTS T-100 Market & Segment http://apps.bts.gov/xml/air_traffic/src/index.xml#MonthlySystem
[9] FAA Aerospace Forecast FY2013-2033
http://www.faa.gov/about/office_org/headquarters_offices/apl/aviation_forecasts/aerospace_forecasts/20132033/media/2013_Forecast.pdf
[10] Inflation CPI Source http://www.usinflationcalculator.com/
[11] US Pilot Labor Supply
http://www.faa.gov/news/conferences_events/aviation_forecast_2010/agenda/media/GAF%20Jim%20Higgins%20and%20Kent
%20Love.pdf
58
[12] Liu & Wu: Modeling Psychological Refractory Period (PRP) and Practice Effect on PRP with
Queuing Networks and Reinforcement Learning Algorithms
[13] Avionics Picture http://www.electronics-cooling.com/1999/09/design-and-reliability-considerations-in-avionics-electronicspackaging/
[14] U2 Cockpit http://www.flightglobal.com/blogs/aircraft-pictures/2008/04/lockheed-u2-cockpit/
[15] NTSB Accident Data: http://www.ntsb.gov/data/aviation_stats.html
[16] Front Page 3 Pilots: http://www.avweb.com/news/ceocockpit/ceo_of_the_cockpit_84_terms_of_up_gearment_198161-1.html
[17] http://www.anotherdeals.com/wp-content/uploads/2014/02/getting-a-job-as-a-commercial-pilot.jpg
59