1. George Mason University March 18, 2004 Harvard University
Air Transportation is a Complex Adaptive System: Not an Air Traffic Control Automation Problem
Dr. George L. Donohue
George Mason University
March 18, 2004 Harvard University
© George Donohue 2004

2. Outline How did We Get Here? Why Should We Care? Capacity - Delay
Capacity – Delay – Safety
System Network Effects
Observations and Recommendations

3. How did We Get Here?
1903 Wright Bros. produced a Heavier than Air Flying System: AGE OF INVENTION
Airfoils, L/W Structure, Controls, L/W Propulsion
WW II ( +50 Yrs) System is Upgraded: AGE OF ALL WEATHER COMMERCIAL FLIGHT
Radar, Jet Aircraft, Radio Navigation and Communication
2003 ( +100 Yrs.) System Needs Upgrading Again: AGE OF RELIABLE INTERNATIONAL TRANSPORTATION NETWORK
Predictable under all Weather Conditions
Maximum Airport Capacity Utilization
Near Optimal Network Load Balancing
Predicable Safety Operating Margins

4. Barriers to the Third Age Vision
The Technical Community has been Aware of the Transition Problem for Over 20 Years!
Increasing Delays, Flight Cancellations
Increasing Runway Incursions, ATC Op Errors and TCAS RA’s
The Technical Elements that will enable the Development of an Affordable, Reliable, and Predictable Mode of International Transportation Already Exist!
TCAS II Deployed Worldwide No System Credit
FMS with ±30 Sec. RTA ? No System Credit
GPS Navigation and Surveillance (ADS-B) No System Credit
Digital Communication Data Links NOT DEPLOYED
TMA, pFAST, aFAST, URET No System Credit
The Barriers to Growth are Regulatory and Institutional!

5. Increasing Delay is a Frog in the Boiling Water Problem
Deregulation
Source: USDOT BTS NTS 2000; USDOT BTS update April, 2002; DOC BEA 2002 (*real GDP using 1996 chained dollars

6. Air Transportation’s Contribution to GDP

7. Civil Transport Share of Aerospace Industry

8. Capacity and Delay
System Capacity is Primarily Limited by Network Runway Availability
ATC Workload is an important Secondary Limitation
Runway Maximum Capacity is a function of Aircraft Landing Speed and Runway Occupancy Time (ROT)
Delay is a Non-Linear function of Demand to Maximum Capacity Ratio
Stochastic FCFS System
Queuing Theory Applies
Major Hub Airports are Over-Scheduled
Transportation Network Is NOT LOAD BALANCED
Market Mechanisms Could Achieve this Goal

9. Network Operational Capacity is a Limited Commodity
CMAX = 2 C AR MAX S i (XG)i Ri {Airports}
–K AK(t) {Airspace Management Intervention}
S = f ( Safety, ATC , Wake Vortex, etc.) ~ 0.6 to 0.8
AK(t) = (A/CREQUEST – A/CACCEPT) ~ [ 0 to >1,000]
AK(t) = f ( GDP:Weather, Sector Workload Constraints )
C AR MAX ~ 40 Arrivals/Hour (set by Runway Occupancy Time)
Ri = Number of Runways at ith Airport
XGi = Airport Configuration Factor at ith Airport
i = 1 to N, where N is approximately 60 Airports
K = 1 to M, where M is typically much less than 100 Sectors

10. ATS Delays Grow Exponentially with Increasing Capacity Fraction

11. Aircraft Arrival Rate: Distance-Time Relationship
Spacing
(sec) 60
ROT? 72 90
WV?
120
180

12. NY LaGuardia: A non-Hub Maximum Capacity Airport
1 Arrival Runway
1 Departure Runway
45 Arrivals/Hr (Max)
80 Seconds Between Arrivals
11.3 minute Average Delay
77 Delays/1000 Operations
40 min./Delay
Airport is overscheduled almost entire day (peak one op/min)

13. New York LaGuardia Airport Arrival- Departure Spacing VMC
Data provides inputs to airspace models
Significant observations of operations outside of calculated capacity, possible hazard
45 Arr./Hr/RW
@ 80 sec separation
DoT/FAA

14. LGA Arrival - Departure IMC

15. Capacity-Delay-Safety
ATM System Safety and Capacity are Non-Linearly Related
Wake Vortex Separation sets the Current System Capacity Limit in Instrument Meteorological Conditions
Safety Limitation
ICAO System Safety Goal is 10-9 / Operation
Small number Statistics leads us to use Accident Precursors as Safety Indicators
Safety Analysis must be Analytical

16. Time Separation Is an Important Determinant of the SAFETY Limitation
35+ Arrivals/RW/Hr
ROT
Probability
A/C Inter-arrival Time
Time (seconds)

17. Accident Pre-Cursor Incidents Safety – Capacity Relationship
ATL, BWI, LGA, DCA
Haynie, GMU 2002

18. ATL Estimated Collision Probability
Collision probability per SRO for each combination
Small-Heavy
Leader - Trailer
Small-Large
Small-B757
Using TOPAZ collision risk model, we calculated the collision probability per simultaneous runway occupancy for each aircraft combination. It’s shown that the most possible collision case is that a heavy aircraft follows a small aircraft if they are involved in a simultaneous runway occupancy. The cases of a large aircraft following a small one and a B757 following a small are of the next top possibilities.
Richard Y. Xie, GMU research in progress

19. Wake Vortex Accident Rate in Safety-Capacity Coordinates
30 Years
3 Years
NLR Stochastic Analysis

20. System Network Effects
Aprox. 10 Major Hub Airports are Operating at D/C max > 0.65
Delays at these Airports spread Non-Linearly throughout the Network
Runway Additions at one Airport May have Little Network Effect
System-wide improvements have a Larger Effect than Individual Airport Improvements
Except at Major Airports like ORD, LGA and ATL!

21. Major US Airport Congestion
Queuing Delays Grow Rapidly
J. D. Welch and R.T. Lloyd, ATM 2001

22. Source: William DeCota, Port Authority of New York
Market Does Not Act to Minimize Delay or Maintain SAFETY: LGA Air 21 Impact
Source: William DeCota, Port Authority of New York

23. Atlanta: A Maximum Capacity Fortress Hub Airport
2 Runways – Arrivals
2 Runways – Departures
50 Arrivals/Hr/RW – Max
72 Seconds Between Arrivals
8.5 minutes Average Delay
36 Delays/1000 Operations
38 min./delay

24. Simulated Auction Delay Benefit at ATL
45 min maximum schedule deviation allowed,
no flights are rerouted
Scheduled arrivals (#operations/quarter hour)
Estimated Average Runway Queuing Delay (min)
ATL reported optimum rate
Loan Le Research in Progress

25. Observations – NAS Safety
We are approaching the Point that the existing system may be demonstrably less safe (at current and future capacity fractions) than a new, more synchronous, aircraft FMS/ADS-B separation based system
System is Safe BUT Safety Margins are Diminishing!
This case has not been Analyzed nor even Suggested as a RATIONAL for CHANGE to date!

26. Central Research Questions
Both Safety and Efficiency Concerns lead us to the conclusion that the network should be operated as a Synchronous System
with economic incentives to use the largest aircraft affordable and economically viable
Time Window Auctions at Airport Metering Fix may provide the Economic Incentives necessary to maximize/optimize Network Capacity
Central Research Questions:
How Synchronous Can We Make this System – All Weather?
What Should the Design Target Level of Safety Be?

27. Backup Slides

28. Wake Normalized Aircraft Time Separation: LGA in VMC & IMC

29. Observed WV Separation Violations vs. Capacity Ratio
Haynie, GMU 2002

30. ATL Arrival - Departure IMC

31. ATL and LGA Inter-Arrival Time in IMC and VMC:32 - 39 Ar/Rw/Hr

32. Observations - Recommendations
FAA Culture – Barriers to Change
NASA Culture – Barriers to Change
State of NAS Safety
Proposed Grand Experiment/OPEVAL

33. FAA Barriers to Change FAA has an Operational and Regulatory Culture
Inclination to follow training that has seemed to be Safe in the Past
FAR has NOT Changed to Provide Operational Benefits from Introduction of New Technology
Assumption that Aircraft Equipage would be Benefits Driven did not account for Lack of an ECONOMIC and/or SAFETY Bootstrapping Requirement

34. FAA Investment Analysis Primarily focus on Capacity and Delay
OMB requirement to have a B/C ratio > 1 leads to a modernization emphasis on Decreasing Delay
In an Asynchronous Transportation Network operating near it’s capacity margin, Delay is Inevitable
Delay Costs Airlines Money and is an Annoyance to Passengers BUT
is Usually Politically and Socially Acceptable

35. NASA Barriers to Change
NASA has become more Process Oriented than Product Oriented
Frequently Stated Objective of 25 year Implementation Goal Avoids Accountability and renders NASA TRL 4/6 Product unusable by either Government Agencies or Industry
NASA needs a Cadre of Engineers/System Analysts with a long range goal of becoming the USG Technical Experts in Aviation System Safety Analysis

36. Proposed Grand Experiment: OPEVAL to FOCUS Efforts
FY 2008 One Year of Night Operations
12pm to 8 am
DAG-TM + aFAST+CDM + WV
Entire US Air Cargo Fleet
Inter-Agency IPT
DoT, NASA, FAA, DoD, NTSB, Boeing, CAA, Airlines

37. Hypothesis: Most Major Changes to the NAS have been due to Safety Concerns
1960’s Mandated Introduction of Radar Separation
1970’s Decrease in Oceanic Separation Standards Required a Landmark Safety Analysis
1970’s Required A/C Transponder Equipage
1970’s Required A/C Ground Proximity Equipage
1990’s Required A/C TCAS Equipage
1990’s Required A/C Enhanced Ground Prox. Equipage
1990’s TDWR & ITWS Introduction
1990’s Mandated Development of GPS/WAAS