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 Project Management 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 Flight safety is maintained 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 200.00 Operating Expense (Billions) 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 [5] & [6] 180.00 160.00 140.00 120.00 100.00 80.00 60.00 40.00 20.00 0.00 Industry Profitable Cost Per Pilot 2012 2011 2010 2009 2008 2007 Cost Increasing 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991 Industry In Decline & Salary Negotiation 1990 Cost (Thousands) Cost Per Pilot Year The cost per pilot has a period of decline from 2000-2008 Slowly increasing at present Pay schedules established for salary increases [6] BTS Schedule P10 [7] BTS Schedule P5.2 *Values are inflation adjusted to 2012 based on consumer price index 9 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 10 Radio Flight Navigator Co-Pilot? Operator Engineer Is a Single Pilot Cockpit the Next Step? 11 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 12 Cockpit Avionics MCP Center Instrument Panel 1st Officer’s Instrument Panel MCDU Overhead Panel 13 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) 14 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 15 Example Procedure 16 Action Frequencies 160 140 120 100 80 PF 60 PNF 40 B/P 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% 17 17 Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management 18 Stakeholder Interactions Support SPC Reserved about SPC Oppose SPC 19 19 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 2020 (FAA, DoT) Aviation Workforce (Pilots, Pilots’ Unions, ATC, ATC Unions) Customer Base Aviation Industry (Air Carriers, Management, Manufactures, Insurance, & Airports) Tension with SPC Agenda Background & Context Stakeholder Analysis Problem & Need Design Alternatives Simulation & Methodology Results Recommendation Project Management 21 Problem Statement Rising operating expense contributes to financial instability in commercial aviation Profitability is difficult to achieve Pilot labor shortage predicted 22 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 23 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 24 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 +10-15 Years +20-30 Years Two Pilot Cockpit Two Pilot Cockpit with Alternative Evaluation of Alternative +40 Years Single Pilot Cockpit 25 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. 26 Regulatory Agencies Aviation Workforce Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management 27 Requirements Requirement 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 Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management 29 Design Alternatives 1. 2. 3. 4. Two Pilot Cockpit (No Change) Single Pilot with No Support Onboard Procedure Support System Remote Ground Pilot Terminal 30 Physical Process Diagram Flight Goal Start Aircraft State System Target Aircraft State Pilots Procedures 31 Integration Concept MCP Center Instrument Panel Remote Alternative 1st Officer’s Instrument Panel Onboard Alternative MCDU Overhead Panel 32 Baseline Two Pilot Cockpit Single Pilot No Automation Procedure Support System Ground Pilot Terminal No Change to current two pilot system Nothing is modified from operating manual 33 Baseline Two Pilot Cockpit Single Pilot No Procedure Support Procedure Support System Ground Pilot Terminal Remove the co-pilot and transfer all the procedures to the remaining pilot No new systems supporting the single pilot 34 Baseline Two Pilot Cockpit Single Pilot No Procedure Support Procedure Support System Ground Pilot Terminal All PNF tasks moved to the PF Eliminate callouts 35 Baseline Two Pilot Cockpit Single Pilot No Procedure Support Procedure Support System Ground Pilot Terminal 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 36 Baseline Two Pilot Cockpit Single Pilot No Procedure Support Some PNF actions taken over by automation Used for situational awareness for PF during a procedure Procedure Support System Ground Pilot Terminal 37 Baseline Two Pilot Cockpit Single Pilot No Procedure Support Procedure Support System Ground Pilot Terminal Ground pilot monitors flight and interacts with ATC on behalf of the airplane No control capability Requires specialized ground networks and infrastructure 38 Baseline Two Pilot Cockpit Single Pilot No Procedure Support Procedure Support System Ground Pilot Terminal Some PNF actions taken over by ground pilot Primarily handles ATC & diagnostic procedures 39 Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management 40 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 41 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 42 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 Procedure Model Simulation Processing Time 43 Procedural Simulation Formulas Simulation 𝑙 𝑘 𝐴𝑃𝑇 = 𝒘𝑷 = 𝑃𝑖 = 𝑤𝑖 𝑃𝑖 𝑜 𝑇𝑗 𝑇𝑗 = 𝑗=1 𝑖=1 𝐴𝑚 𝑚=1 𝐴𝑚 ~ln(𝑁 𝜇, 𝜎 2 ) 𝐴𝑚 ~U(a, b) 𝐴𝑚 ~Tri(c, d, e) 𝐴𝑃𝑇 = 𝑎𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 𝑝𝑟𝑜𝑐𝑒𝑠𝑠𝑖𝑛𝑔 𝑡𝑖𝑚𝑒 𝑃𝑖 = 𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑒 𝑇𝑗 = 𝑡𝑎𝑠𝑘 𝐴𝑚 = 𝑎𝑐𝑡𝑖𝑜𝑛 𝑤𝑖 = 𝑝𝑟𝑜𝑐𝑒𝑑𝑢𝑟𝑎𝑙 𝑤𝑒𝑖𝑔ℎ𝑡 Replications 𝑍𝛼 ∗ 𝑆0 𝑅𝑒𝑝𝑙𝑖𝑐𝑎𝑡𝑖𝑜𝑛𝑠 ≥ 2 𝜀 Analysis 2 𝑡= 𝐴𝑃𝑇𝐶𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐴𝑃𝑇𝐴𝑙𝑡 𝑠12 𝑠22 + 𝑁1 𝑁2 𝐻0 : 𝐴𝑃𝑇𝐶𝑜𝑛𝑡𝑟𝑜𝑙 ≥ 𝐴𝑃𝑇𝐴𝑙𝑡 𝐻1 : 𝐴𝑃𝑇𝐶𝑜𝑛𝑡𝑟𝑜𝑙 < 𝐴𝑃𝑇𝐴𝑙𝑡 𝑅𝑒𝑗𝑒𝑐𝑡 𝑖𝑓: 𝑃 > 0.05 44 44 Design of Experiment Input Output Alternative Procedures Tasks Action Times Alternative Processing Times Single Pilot No Support P1…Pn T1m…Tnm Lognormal~A1r…Amr TBD Two Pilot P1…Pn T1o…Tno Lognormal~A1r…Aor TBD Procedure Support System P1…Pn T1p…Tnp Lognormal~A1r…Apr Uniform~A1r…Apr TBD Ground Pilot Terminal P1…Pn T1q…Tnq Lognormal~A1r…Aqr Triangular~A1r…Aqr TBD 45 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 46 Business Case Formulas Cost Formulas 𝑁 𝐸 𝐴𝐿𝐶 = 𝐶𝐴𝑙𝑡 + 𝑡=1 𝐶𝑃𝑖𝑙𝑜𝑡 + 1+𝑑 𝑡 𝐶𝐴𝑙𝑡 = 𝐴𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑖𝑣𝑒 𝐶𝑜𝑠𝑡 𝐶𝑂𝑡ℎ𝐿𝑎𝑏𝑜𝑟 = 𝑂𝑡ℎ𝑒𝑟 𝐿𝑎𝑏𝑜𝑟 𝐶𝑜𝑠𝑡𝑠 𝑁 𝑡=1 𝐶𝑂𝑡ℎ𝐿𝑎𝑏𝑜𝑟 + 1+𝑑 𝑡 𝑁 𝑡=1 𝐶𝑃𝑖𝑙𝑜𝑡 = 𝑃𝑖𝑙𝑜𝑡 𝐿𝑎𝑏𝑜𝑟 𝐶𝑜𝑠𝑡 𝐶𝑀𝑎𝑖𝑛𝑡 1+𝑑 𝑡 𝑑 = 𝐼𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝐶𝑀𝑎𝑖𝑛𝑡 = 𝐴𝑖𝑟𝑐𝑟𝑎𝑓𝑡 𝑀𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒 𝐶𝑜𝑠𝑡𝑠 47 Design of Experiment Input 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) Procedure Support System Triangular(d,e,f) Triangular(a,b,c) Triangular(a,b,c) Ground Pilot Terminal Triangular(g,h,i) Triangular(a,b,c) Triangular(a,b,c) 48 Value Hierarchy Single Pilot Utility Alternative Processing Time In Scope Out of Scope Safety TRL Profitability Training Liability Maintenance Usability 49 Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management 50 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 51 Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management 52 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 Today +10-15 Years +20-30 Years Two Pilot Cockpit Two Pilot Cockpit with Alternative Evaluation of Alternative +40 Years Single Pilot Cockpit 53 Agenda Background & Context Stakeholder Analysis Problem & Need Requirements Design Alternatives Simulation & Methodology Results Recommendation Project Management 54 Work Breakdown 55 Schedule 56 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 Risk Risk Simulation Complexity Procedure Complexity Description 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. The number of procedures to be modeled is very large and takes significant resources due to manual nature of input. Mitigation Plan to devote significant work to simulation during the winter break period. Simulation coding started ahead of schedule. Plan to make baseline model and make simplifications to existing structure rather than re-authoring all procedures. Busy Co-Sponsor NASA Ames researcher has been unavailable to support project. Working with other sponsors to get additional resources. 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. Rating Likelihood: Likely Impact: Major Likelihood: Likely Impact: Major Likelihood: Likely Impact: Minor Likelihood: Very Likely Impact: Major 58 Budgeting Parameters $50.00 base hourly rate $106.38 hourly rate w/ GMU overhead factor Estimate 701 hours Budgeted Cost of $74,574.47 59 EVM $110,000.00 $90,000.00 $70,000.00 Faculty Presentations Dollars EV $50,000.00 PV Final Draft AC Conferences $30,000.00 Abstract & Manuscript Worst Case Best Case Final Proposal $10,000.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 -$10,000.00 Week 60 CPI / SPI 1.40 1.20 1.00 Ratio 0.80 CPI 0.60 SPI 0.40 0.20 0.00 1 2 3 4 5 6 Week 7 8 9 10 61 Questions? 62 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 63 Backup 64 [7] Projected Passenger Miles Passenger Miles (Billions) 1,200 1,000 800 600 Domestic 400 Total 200 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 0 Year Passenger demand increase of 30% projected over next 10 years [8] Commercial aviation under pressure to meet demand while achieving profitability [8] BTS T-100 Market & Segment [9] FAA Aerospace Forecast FY2013-2033 *Values are inflation adjusted to 2012 based on consumer price index 65 65 Cost Model Input Description Average Fleet Age The average fleet age as timeframe for lifecycle of “typical” aircraft Average Pilot Cost Yearly pilot cost based on average pay schedules with pay increase Alternative Investment The expected cost per unit to deploy alternative Interest Rate of forgoing alternative investments Other Aircraft Operating Expenses Operating expense for aircraft as reported through BTS treated as random variates 66 Alternatives Satisfying Requirements Mission Requirement Single Pilot No Support Two Pilots Task Support System Ground Pilot Terminal Task Load X TBD TBD Reliability X TBD Operating Expense TBD TBD TBD Lifecycle Cost TBD TBD TBD 67 Decompose tasks into Excel Select roles and cognitive functions Visual Instrument Inspection Visual Environment Inspection Verbal Callout Physical Instrument Manipulation Etc. 68 68 Parse Excel spreadsheet into an XML schema Hierarchy of tasks captured by XML Parsed into simulation 69 69