Design of a Single Pilot Cockpit for Airline Operations

DESIGN OF A SINGLE PILOT COCKPIT FOR AIRLINE OPERATIONSJonathan Graham, Chris Hopkins, Andrew Loeber, Soham Trivedi

2

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

3

Agenda Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management

Scope Large Commercial Air Transportation

Passenger and Cargo Carriers Carriers With Operating Revenue >$20 Million Domestic Operations

US Airspace National Airspace System (NAS) ATC FAA Regulatory Body

Financial data adjusted for inflation to 2012 dollars

Flight safety is maintained4

5

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

-40,000,000,000.00

-30,000,000,000.00

-20,000,000,000.00

-10,000,000,000.00

0.00

10,000,000,000.00

20,000,000,000.00

Profit/Loss for Major Air Carriers

Profit/LossNet Income

Year

Dol

lars

(Bi

llion

s)

[1]

Bankruptcies Filed

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

6

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

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

0.0020,000,000,000.0040,000,000,000.0060,000,000,000.0080,000,000,000.00

100,000,000,000.00120,000,000,000.00140,000,000,000.00160,000,000,000.00180,000,000,000.00200,000,000,000.00

Projected Total Operating Expense for Major Air Carriers

Year

Ope

ratin

g Ex

pens

e (B

illio

ns)

Total Op-erating Expense

Projected Total Op-erating Expense

[2]

7

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

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

0.0010.0020.0030.0040.0050.0060.00

% Operations Expense for Major Air Carriers

% Pilot% Fuel% Dir Maint% Air Ops

Year

% O

pera

tions

Cos

t[3]

[3] BTS Schedule P5.2*Values are inflation adjusted to 2012 based on consumer price index

8

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]

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

01000020000300004000050000600007000080000

Projected Pilots

Year

Pilo

ts (

Thou

sand

s)[4]

Total Pilots

Projected Supply

Projected Pilot De-mand

9

The cost per pilot has a period of decline from 2000-2008 Slowly increasing at present Pay schedules established for salary increases

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

0.0020,000.0040,000.0060,000.0080,000.00

100,000.00120,000.00140,000.00160,000.00180,000.00 Cost Per Pilot

Cost Per Pilot

Year

Cost

(Th

ousa

nds)

[5] & [6]

Industry Prof-itable Industry In De-

cline & Salary Negotiation

Cost Increasing

[6] BTS Schedule P10[7] BTS Schedule P5.2*Values are inflation adjusted to 2012 based on consumer price index

10

Historical Cockpit Devolution Historically the need for cost reduction

has changed the roles of the flight crew Size of the cockpit flight crew has been

shrinking

11

Navigator

Radio Operator

Flight EngineerCo-Pilot?

Is a Single Pilot Cockpit the Next Step?

12

The Flight Deck Two pilots operate the aircraft

Pilot Flying Flies the Aircraft Confirms Callouts

Pilot Not Flying 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

13

Cockpit Avionics1st Officer’s Instrument Panel

Center Instrument Panel

MCP

Overhead PanelMCDU

14

Flight Procedures Procedures are followed to operate an aircraft Flight Crew Operating Manual (FCOM) RJ100

Describes flight procedures Official FAA approved document

Identifies responsible entities for procedural tasks Pilot Flying (PF) Pilot Not Flying (PNF)

15

Procedure

Tasks

Actions

A procedure is decomposed by tasks and physical/cognitive actions that are required to execute a task

FCOM Codifies Procedures 63 Procedures 613 Total Tasks 737 Total Actions

Procedures are decomposed to identify potential reallocation of actions to a design alternative

Procedure Decomposition

16

Example Procedure

1717

04080

120160

Action Frequencies

PFPNFB/P

PF40%

PNF55%

B/P5%

% Actions by Pilot

18

Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management

19

Stakeholder Interactions

19

Support SPC

Reserved about SPC

Oppose SPC

20

StakeholderGroup

PrimaryObjectives Tension with SPC

Regulatory Agencies(FAA, DoT)

• Maximize:• Flight safety• Consumer

protection

A SPC would inherently introduce new risks and decrease overall flight safety, leading regulatory agencies to withhold their approval

Aviation Workforce (Pilots, Pilots’ Unions, ATC, ATC

Unions)

• 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

Customer Base• 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

Aviation Industry(Air Carriers, Management, Manufactures, Insurance, &

Airports)

• 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

20

21

Agenda Background & Context Stakeholder Analysis Problem & Need Design Alternatives Simulation & Methodology Results Recommendation Project Management

22

Problem Statement Rising operating expense contributes to

financial instability in commercial aviation Profitability is difficult to achieve

Pilot labor shortage predicted

23

GAP Analysis

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

0

0.2

0.4

0.6

0.8

1

1.2

1.4Operating Revenue to Operating Expense Ratio

Ra...

Year

Ratio

*BTS Schedule P5.2

*Values are inflation adjusted to 2012 based on consumer price index

Target Ratio

24

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

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

Two Pilot Cockpit

Two Pilot Cockpit

with Alternative

Single Pilot Cockpit

Evaluation of

Alternative 25

Today

+10-15 Years

+20-30 Years

+40 Years

Organization

Tension to be Mitigated Benefit to Slow Phase-In

Regulatory Agencies

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.

Aviation Workforce

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.

26

27


RequirementsRequirement ID

Description

M.1 The single pilot cockpit system shall reduce or maintain the baseline pilot flying procedure time of TBX.1

M.2 The single pilot cockpit system shall meet ARP4761 Level A assurance of 1 failure per billion 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 TBX.2 dollars.

28

29


30

Design Alternatives1. Two Pilot Cockpit (No Change)2. Single Pilot with No Support3. Onboard Procedure Support System4. Remote Ground Pilot Terminal

31

Physical Process Diagram

SystemStart Aircraft State

Procedures

Target Aircraft State

Flight Goal

Pilots

32

Integration Concept

Onboard Alternativ

eRemote

Alternative

1st Officer’s Instrument PanelCenter Instrument Panel

MCP

Overhead PanelMCDU

33

No Change to current two pilot system Nothing is modified from operating

manual

Baseline Two Pilot Cockpit

Single Pilot No Automation

Procedure Support System

Ground Pilot Terminal

34

Remove the co-pilot and transfer all the procedures to the remaining pilot

No new systems supporting the single pilot


Single Pilot No Procedure Support



35





All PNF tasks moved to the PF

Eliminate callouts

36

Onboard system integrating with avionics Can receive inputs from avionics to

automate pilot actions and provide feedback to pilot flying

No ability to control plane Report to PF who controls ops Interactive front end for pilot





37





Some PNF actions taken over by automation

Used for situational awareness for PF during a procedure

38

Ground pilot monitors flight and interacts with ATC on behalf of the airplane

No control capability Requires specialized

ground networks and infrastructure





39





Some PNF actions taken over by ground pilot

Primarily handles ATC & diagnostic procedures

40


41

Model & Simulation Assumptions Flight safety is maintained or improved Keystroke Level Model and other human computer interaction

parameters approximate operator actions in cockpit Ignore future/emerging aviation systems Human factors out of scope 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 Ignore additional company specific pilot tasks

42

Procedural Model 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

Procedural Simulation Java program Input procedural model

for each alternative Output Alternative

Processing Time (APT) Run several simulation

replications for each alternative and perform statistical tests on results

43

SimulationProcedure Model Processing Time

44

AnalysisReplications

Procedural Simulation Formulas

Simulation

𝑃 𝑖=∑𝑗=1

𝑙

𝑇 𝑗𝐴𝑃𝑇=𝒘𝑷=∑𝑖=1

𝑘

𝑤𝑖𝑃 𝑖 𝑇 𝑗=∑𝑚=1

𝑜

𝐴𝑚𝐴𝑚 ln (𝑁 (𝜇 ,𝜎 2 ))

𝑃 𝑖=𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒𝐴𝑃𝑇=𝑎𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒𝑝𝑟𝑜𝑐𝑒𝑠𝑠𝑖𝑛𝑔𝑡𝑖𝑚𝑒 𝑇 𝑗=𝑡𝑎𝑠𝑘

𝐴𝑚 U (a ,b)𝐴𝑚 Tri (c ,d ,e )𝐴𝑚=𝑎𝑐𝑡𝑖𝑜𝑛𝑤𝑖=𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑎𝑙 h𝑤𝑒𝑖𝑔 𝑡

44𝑅𝑒𝑝𝑙𝑖𝑐𝑎𝑡𝑖𝑜𝑛𝑠≥( 𝑍 𝛼

2

∗𝑆0

𝜀 )2 𝑡= 𝐴𝑃𝑇𝐶𝑜𝑛𝑡𝑟𝑜𝑙−𝐴𝑃𝑇 𝐴𝑙𝑡

√ 𝑠12

𝑁 1+𝑠2

2

𝑁2 𝑅𝑒𝑗𝑒𝑐𝑡 𝑖𝑓 :𝑃>0.05𝐻1 : 𝐴𝑃𝑇 𝐶𝑜𝑛𝑡𝑟𝑜𝑙< 𝐴𝑃𝑇 𝐴𝑙𝑡

𝐻0 : 𝐴𝑃𝑇𝐶𝑜𝑛𝑡𝑟𝑜𝑙≥𝐴𝑃𝑇 𝐴𝑙𝑡

Input Output

Alternative

Procedures Tasks Action Times

Alternative Processing

TimesSingle

Pilot No Support

P1…PnT1m…Tnm

Lognormal~A1r…Amr TBD

Two Pilot P1…PnT1o…Tno

Lognormal~A1r…Aor TBD


P1…Pn T1p…TnpLognormal~A1r

…Apr

Uniform~A1r…Apr TBD

Ground Pilot

TerminalP1…Pn

T1q…Tnq

Lognormal~A1r…Aqr

Triangular~A1r…Aqr

TBD

Design of Experiment

45

46

Business Case Develop model and simulation to analyze

alternatives’ cost feasibility Aircraft lifecycle cost

Data collected for Boeing 737-300 (most common aircraft)

Random variables describe costs (BTS data) Use Monte Carlo to get average aircraft lifecycle cost If LCC>Labor Cost – system is not worthwhile

Business Case Formulas

47

Cost Formulas

𝐸 [ 𝐴𝐿𝐶 ]=𝐶𝐴𝑙𝑡+∑𝑡=1

𝑁 𝐶𝑃𝑖𝑙𝑜𝑡

(1+𝑑 )𝑡+∑𝑡=1

𝑁 𝐶 h𝑂𝑡 𝐿𝑎𝑏𝑜𝑟

(1+𝑑 )𝑡+∑𝑡=1

𝑁 𝐶𝑀𝑎𝑖𝑛𝑡

(1+𝑑 )𝑡

𝐶 h𝑂𝑡 𝐿𝑎𝑏𝑜𝑟= h𝑂𝑡 𝑒𝑟 𝐿𝑎𝑏𝑜𝑟 𝐶𝑜𝑠𝑡𝑠𝐶𝑃𝑖𝑙𝑜𝑡=𝑃𝑖𝑙𝑜𝑡 𝐿𝑎𝑏𝑜𝑟 𝐶𝑜𝑠𝑡𝐶𝐴𝑙𝑡=𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒𝐶𝑜𝑠𝑡𝐶𝑀𝑎𝑖𝑛𝑡=𝐴𝑖𝑟𝑐𝑟𝑎𝑓𝑡𝑀𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒𝐶𝑜𝑠𝑡𝑠

𝑑=𝐼𝑛𝑡𝑒𝑟𝑒𝑠𝑡

48

Design of ExperimentInput Output

Alternative Alternative Cost

Other Labor &

Maintenance

Expected Aircraft Lifecycle Cost

Single Pilot No Automation

Triangular(a,b,c)

Triangular(a,b,c) Triangular(a,b,c)

Two Pilot -- Triangular(a,b,c) Triangular(a,b,c)


Triangular(d,e,f)



Triangular(g,h,i)


49

Value HierarchySingle Pilot

Utility

TRLAlternative Processing

TimeSafety Profitabilit

yMaintenan

ce

In ScopeOut of Scope

Training Liability Usability


50

51

Anticipated Results Expect the Procedural Support System and

Ground Pilot Terminal to reduce processing times Unsupported Single Pilot will have highest

processing times Cost of Ground Pilot Terminal will be significant

Cost of Unsupported Single Pilot and Procedure Support will be the least

52


53

Anticipated Recommendation Will 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

Two Pilot Cockpit

Two Pilot Cockpit

with Alternative

Single Pilot Cockpit

Evaluation of

Alternative

Today

+10-15 Years

+20-30 Years

+40 Years

54


55

Work Breakdown

56

Schedule

Schedule Detail Critical Paths:

10.0 Deliverable Preparation & Assembly 8.0 Modeling & Simulation

Majority of subtasks completed in first semester with long run subtasks remaining leading to competition

57

58

RiskRisk Description Mitigation Rating

Simulation Complexity

There is a chance that the complexity of the simulation and task model will cause scheduling delays which may impact the ability to produce

results for the IEEE paper due in February.

Plan to devote significant work to simulation during the winter break period. Simulation

coding started ahead of schedule.

Likelihood: LikelyImpact: Major

Procedure ComplexityThe number of procedures to be modeled is very

large and takes significant resources due to manual nature of input.

Plan to make baseline model and make simplifications to existing structure rather

than re-authoring all procedures.

Likelihood: LikelyImpact: Major

Busy Co-Sponsor NASA Ames researcher has been unavailable to support project.

Working with other sponsors to get additional resources.

Likelihood: LikelyImpact: Minor

Input Data

Serious assumptions were made in regards to task performance modeling. Further information is

required to find additional data to validate assumption or provide actual performance data.

Soliciting feedback from professors, sponsor, and professional pilots. Rely on sensitivity analysis to test assumptions.

Likelihood: Very Likely

Impact: Major

59

Budgeting Parameters $50.00 base hourly rate $106.38 hourly rate w/ GMU overhead

factor Estimate 701 hours Budgeted Cost of $74,574.47

60

1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435363738-$10,000.00

$10,000.00

$30,000.00

$50,000.00

$70,000.00

$90,000.00

$110,000.00

EVM

EVPVACWorst CaseBest Case

Week

Dol

lars

Faculty Presentations

Conferences

Final Proposal

Final Draft

Abstract &

Manuscript

61

1 2 3 4 5 6 7 8 9 100.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

CPI / SPI

CPISPI

Week

Ratio

62

Questions?

63

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

[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/2013-2033/media/2013_Forecast.pdf

[10] Inflation CPI Source http://www.usinflationcalculator.com/

[5] WSJ Airlines Face Acute Shortage of Pilots http://online.wsj.com/news/articles/SB10001424052970203937004578079391643223634#articleTabs%3Darticle

[11] US Pilot Labor Supply http://www.faa.gov/news/conferences_events/aviation_forecast_2010/agenda/media/GAF%20Jim%20Higgins%20and%20Kent%20Love.pdf

64

Backup

65

Passenger demand increase of 30% projected over next 10 years [8]

Commercial aviation under pressure to meet demand while achieving profitability

65

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020

2021

2022

0200,000,000400,000,000600,000,000800,000,000

1,000,000,0001,200,000,000

Projected Passenger Miles

DomesticTotal

Year

Pass

enge

r M

iles

(Bill

ions

)[7]

[8] BTS T-100 Market & Segment[9] FAA Aerospace Forecast FY2013-2033*Values are inflation adjusted to 2012 based on consumer price index

Cost Model

66

Input DescriptionAverage Fleet Age The average fleet age as timeframe for lifecycle

of “typical” aircraftAverage Pilot Cost Yearly pilot cost based on average pay

schedules with pay increaseAlternative Investment

The expected cost per unit to deploy alternative

Interest Rate of forgoing alternative investmentsOther Aircraft Operating Expenses

Operating expense for aircraft as reported through BTS treated as random variates

Alternatives Satisfying Requirements

Mission Requirement

Single Pilot No Support

Two Pilots Task Support System


Task Load X TBD TBDReliability X TBDOperating Expense TBD TBD TBD

Lifecycle Cost

TBD TBD TBD

67

6868

Decompose tasks into Excel Select roles and cognitive functions

Visual Instrument Inspection Visual Environment Inspection Verbal Callout Physical Instrument Manipulation Etc.

6969

Parse Excel spreadsheet into an XML schema

Hierarchy of tasks captured by XML

Parsed into simulation

Design of a Single Pilot Cockpit for Airline Operations

Documents