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

Air Transportation is a Complex Air Transportation is a Complex Adaptive System:Adaptive System:

Not an Air Traffic Control Not an Air Traffic Control Automation ProblemAutomation Problem

Dr. George L. DonohueGeorge Mason University

March 18, 2004 Harvard University© George Donohue 2004

OutlineOutline

• How did We Get Here?

• Why Should We Care?

• Capacity - Delay

• Capacity – Delay – Safety

• System Network Effects

• Observations and Recommendations

How did We Get Here?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

Barriers to the Third Age VisionBarriers 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!

Source: USDOT BTS NTS 2000; USDOT BTS update April, 2002; DOC BEA 2002 (*real GDP using 1996 chained dollars

Increasing Delay is a Frog in the Boiling Water Problem

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

1960 1965 1970 1975 1980 1985 1990 1995 1998

Year

Gro

wth

rat

e/G

DP

gro

wth

rat

e

Air carrierpassenger miles

Highway trip miles

Rail passenger miles

Deregulation

Air Transportation’s Air Transportation’s Contribution to GDPContribution to GDP

Total Output GDP ContributionAir Transportation $205 $80Aircraft Manufacturing $134 $94Tourism $94 $85Agents/Forwarders $3 N/CGovernment $2 N/CTotal Impact $438 $259

Economic Impact of Aviation Industry($billions 1999)

Civil Transport Share of Civil Transport Share of Aerospace IndustryAerospace Industry

$473

$100

$81

$65

$54$31 $6

Source: L. Anderson, op cit., from Forecast International

Large Civil

Transports

Fighter, Attack

and Trainer

Rotorcraft

(Military and

Civil)

Business /

Corporate

Regional /

Commuter

Military

Transport

General

Aviation

Total Projected Aircraft Market 1999 to 2008: $810 Billion

FUTURE MARKETS FOR AERONAUTICS PRODUCTS ARE LARGE

74% CIVIL TRANSPORT

Capacity and DelayCapacity 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

Network Operational Capacity is Network Operational Capacity is a Limited Commoditya 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

ATS Delays Grow Exponentially ATS Delays Grow Exponentially with Increasing Capacity Fractionwith Increasing Capacity Fraction

PREDICTED DELAY vs. CAPACITY FRACTION

0.05.0

10.015.020.025.030.035.040.045.050.0

0% 20% 40% 60% 80% 100%

AIRPORT CAPACITY FRACTION

NU

MB

ER

DE

LA

YS

( >

15 m

in /

1000

ope

rati

ons)

Aircraft Arrival Rate:Aircraft Arrival Rate:Distance-Time RelationshipDistance-Time Relationship

0

10

20

30

40

50

60

70

80

0 1 2 3 4 5 6 7

DISTANCE ( NMi)

AR

RIV

AL

S /

RW

/ H

R 120 Knots

130 Knots

140 Knots

Spacing

(sec)

60

90

180

120

72ROT?

WV?

NY LaGuardia: A non-Hub NY LaGuardia: A non-Hub Maximum Capacity AirportMaximum 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./DelayTOTAL SCHEDULED OPERATIONS AND CURRENT OPTIMUM RATE BOUNDARIES

0

4

8

12

16

20

24

28

32

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Schedule Facility Est. Model Est.

New York LaGuardia Airport New York LaGuardia Airport Arrival- Departure Spacing VMCArrival- Departure Spacing VMC

0

10

20

30

40

50

60

0 10 20 30 40 50 60

Departures per Hour

Arr

ival

s pe

r H

our

ASPM - Apr 2000 - Visual Approaches

ASPM - Oct 2000 - Visual Approaches

Calculated VMC Capacity

Optimum Rate (LGA)

40,40

Each dot represents one hour of actual traffic

during April or October 2000

45 Arr./Hr/RW@ 80 sec separation

DoT/FAA

LGA Arrival - Departure IMCLGA Arrival - Departure IMC

0

10

20

30

40

50

60

0 10 20 30 40 50 60

Departures per Hour

Arr

ival

s p

er H

ou

r

ASPM - April 2000 - Instrument Approaches

ASPM - October 2000 - Instrument Approaches

Calculated IMC Capacity

Reduced Rate (LGA)

32,32

Capacity-Delay-SafetyCapacity-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

Time Separation Is an Important Time Separation Is an Important Determinant of the SAFETY LimitationDeterminant of the SAFETY Limitation

Time (seconds)

Probability

ROT

A/C Inter-arrival Time

35+ Arrivals/RW/Hr

Accident Pre-Cursor Incidents Accident Pre-Cursor Incidents Safety – Capacity RelationshipSafety – Capacity Relationship

Haynie, GMU 2002

Hazard Reports 1988-2001

0

20

40

60

80

100

120

35 40 45 50 55 60 65 70

Percentage Capacity Used

Nu

mb

er R

epo

rts

Fil

ed

NMAC

RWY Inc

Legal Sep

ATL, BWI, LGA, DCA

ATL Estimated ATL Estimated Collision ProbabilityCollision Probability

Collision probability per SRO for each combination

Small-Heavy

Small-Large

Small-B757

Leader - Trailer

Richard Y. Xie, GMU research in progressRichard Y. Xie, GMU research in progress

Wake Vortex Accident Rate in Wake Vortex Accident Rate in Safety-Capacity CoordinatesSafety-Capacity Coordinates

Single Runway Estimated Wake Vortex Accident Rate50% Mix B747 & B737: S-Wake Calculation

y = -25004Ln(x) + 593477

R2 = 0.8731

y = -37168Ln(x) + 832913

R2 = 0.9698

-

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

450,000

- 1,000,000 2,000,000 3,000,000 4,000,000 5,000,000 6,000,000 7,000,000Safety - Arrivals / WV Accident

Cap

aci

ty

- A

rriv

als

/ Y

ear Log. (Hazardous Accident)

Log. (Catastrophic Accident)

NLR Stochastic AnalysisNLR Stochastic Analysis

30 Years3 Years

System Network EffectsSystem 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!

Major US Airport CongestionMajor US Airport Congestion

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

DEN

BWI

JFK

PIT

CLT

DFW

DTW

IAD

EWR

PHL

SFO

LGA

MSP

SEA

ORD

STL

ATL

LAXA

IRP

OR

T

DEMAND / CAPACITY RATIO

J. D. Welch and R.T. Lloyd, ATM 2001

Queuing Delays Grow Rapidly

Market Does Not Act to Minimize Delay or Market Does Not Act to Minimize Delay or Maintain SAFETY: LGA Air 21 ImpactMaintain SAFETY: LGA Air 21 Impact

Source: William DeCota, Port Authority of New York

LaGuardia Airport

0

2040

60

80

100120

140

160180

200

06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Time of Day

Historic Movements AIR-21 Induced Svc.

Maximum Hourly Operations Based on Current Airspace & ATC Design

Atlanta: A Maximum Capacity Atlanta: A Maximum Capacity Fortress Hub AirportFortress 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

TOTAL SCHEDULED OPERATIONS AND CURRENT OPTIMUM RATE BOUNDARIES

0

10

20

30

40

50

60

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Schedule Facility Est. Model Est.

Simulated AuctionSimulated Auction Delay Benefit at ATL Delay Benefit at ATL

45 min maximum schedule deviation allowed, no flights are rerouted

0

10

20

30

40

50

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

0

5

10

15

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Time (15-min bins)

Original Schedule Auctioned Schedule

Scheduled arrivals (#operations/quarter hour)

Estimated Average Runway Queuing Delay (min)

ATL reported optimum rate

Loan Le Research Loan Le Research in Progressin Progress

Observations – NAS SafetyObservations – 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!

Central Research QuestionsCentral 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?

• Backup Slides

Wake Normalized Aircraft Time Wake Normalized Aircraft Time Separation: LGA in VMC & IMCSeparation: LGA in VMC & IMC

02468

10

-60

-20 20 60 100

140

Seconds Deviation per Aircraft From Perfect WVSS Adherence Value

Air

craf

t / R

W /

Hr (

20 S

ec.

Bin

s)

VFR 33.8 Arr/hrIFR 34 Arr/HrVFR 30.9 Arr/HrVFR 27 Arr/Hr

Observed WV Separation Observed WV Separation Violations vs. Capacity RatioViolations vs. Capacity Ratio

Figure 6-5 Ratio of Incidents to Capacity Used

0

2

4

6

8

0 50 100 150 200

Percent of Capacity Used in 15 Minutes

Nu

mb

er

of

< W

VS

S

Inc

ide

nts

Ex

pe

cte

d i

n

15

Min

ute

s

BWI LGA Quadratic Model

Haynie, GMU 2002

ATL Arrival - Departure IMCATL Arrival - Departure IMC

0

20

40

60

80

100

120

0 20 40 60 80 100 120

Departures per Hour

Arr

iva

ls p

er

Ho

ur

ASPM - April 2000 - Instrument Approaches

Calculated IMC Capacity

Reduced Rate (ATL)

84,90

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

LGA & ATL Arrival Histograms

-2

0

2

4

6

8

10

12

14

0 50 100 150 200 250

Inter-Arrival Time (Seconds)

Air

craf

t / R

W /

Hr

(20

Sec

. Bin

s)

LGA in VMC N=168

LGA in IMC N=124

ATL IN VMC N=114

ATL in VMC N=323

Observations - RecommendationsObservations - Recommendations

• FAA Culture – Barriers to Change

• NASA Culture – Barriers to Change

• State of NAS Safety

• Proposed Grand Experiment/OPEVAL

FAA Barriers to ChangeFAA 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

FAA Investment Analysis Primarily FAA Investment Analysis Primarily focus on Capacity and Delayfocus 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

NASA Barriers to ChangeNASA 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

Proposed Grand Experiment:Proposed Grand Experiment:OPEVAL to FOCUS EffortsOPEVAL 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

Hypothesis: Most Major Changes to the Hypothesis: Most Major Changes to the NAS have been due to NAS have been due to Safety ConcernsSafety 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