Autonomous Formation Flight
MIT Course 16.886, Spring 2004
Air Transportation Systems Architecting
Greg Larson
Program Manager
Boeing Phantom Works
Gerard Schkolnik
Program Manager
NASA DFRC
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 1
Overview
Autonomous Formation Flight: NASA RevCo Program
Boeing is currently engaged with NASA Dryden Flight Research Center on a technically
ambitious project, Autonomous Formation Flight (AFF). The project’s primary goal is to
investigate potential benefits of flying aircraft in the aerodynamic wake vortex emanating
from a lead aircraft’s wing tip. Initial analytic studies predict that a trailing aircraft may
experience drag reductions of 10% or more by gaining additional lift in the updraft portion of
the lead’s wake vortex. The technical challenge is to be able to find the optimal position
within the vortex to fly, then hold that position consistently in what is an extremely turbulent
flow field. We know that pilots have been able to do this in the past, but the task involves a
very high workload.
The Autonomous Formation Flight system marries an extremely robust flight control and
guidance system with a close-coupled GPS/IMU placed on two F-18s. Inter-ship
communication allows the multiple GPS/IMU systems to share state data and through and
extended Kalman filter technique, they yield a differential carrier phase solution. They
resolve the relative position accuracy between the aircraft in formation to less than 10 cm.
Through shared state data, the guidance systems aboard both F-18s resolve coordinated
trajectories that permit the aircraft to maintain formation. The trailing aircraft is thus capable
of maintaining its position within the lead aircraft’s wing tip vortex with extremely high
accuracy.
The implications and applications of this technology are far reaching, not just for fuel
economy but for other future applications such as aerial refueling, aircraft logistics, air traffic
control, and carrier landing systems.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 2
Special Acknowledgements & References To Technical Papers
Jake Vachon (NASA TM 2003-2107341)
Ronald Ray (NASA TM 2003-2107341)
Kevin Walsh (NASA TM 2003-2107341)
Kimberly Ennix (NASA TM 2003-2107341)
Ron Ray (NASA TM 2002 210723)
Brent Cobleigh (NASA TM 2002 210723)
Jake Vachon (NASA TM 2002 210723)
Clint St. John (NASA TM 2002 210723)
Eugene Lavretsky (AIAA-2002-4757)
Glenn Beaver (NASA TM-2002-210728)
Peter Urschel (NASA TM-2002-210728)
Curtis E. Hanson (NASA TM-2002-210728, NASA TM-2002-210729)
Jennifer Hanson (AIAA-2002-3432)
Jack Ryan (NASA TM-2002-210729)
Michael J. Allen (NASA TM-2002-210729)
Steven R. Jacobson (NASA TM-2002-210729)
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 3
Presentation Outline
•
•
•
•
•
•
•
Project Summary
Objectives
Theory
Experiment Design
Phase 0 Flight Test
Phase 1 Flight Test
Cruise Mission
Demonstration
• Performance Seeking
Control
• Aerial Refueling
• Concluding Remarks
Test flights began in August and culminated with a drag-reduction demonstration flight in the beginning of December 2001.
A total of 28 flights were accomplished, and the full test point matrix was accomplished at both M=0.56, 25000 feet, and M=0.86, 36000 feet.
415 test points were flown
5 Project Pilots were involved in AFF Phase One Risk Reduction
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 4
Autonomous
Autonomous Formation
Formation Flight
Flight
• Background
– Many bird species fly in “V” formation to take advantage of the
up-wash field generated by adjacent birds, resulting in less
energy expended.
– Analytical studies and recent AFF flight tests validate these
observations.
• AFF Objectives
– Validate drag reduction concept and prediction tools of a
system of aircraft in formation in the flight environment
– Develop and evaluate sensor and control methodologies for
autonomous close formation flight
• Approach
– Flight test autonomous station keeping control laws of pair of F-18
aircraft.
– Validate drag benefits and wing tip vortex behavior using piloted
flight tests.
– Develop and validate advanced relative GPS system capable of 10
cm relative position accuracy.
– Integrate updated sensors and advanced formation control laws to
perform autonomous station keeping within the vortex wake of a
lead aircraft.
• Benefits
– Potential commercial fuel savings of $0.5 to 1 million per year per
trailing aircraft.
– Application to UAV Swarming, & Aerial Refueling.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 5
Primary Project Objective:
Demonstrate Drag Reduction
Drag
Drag Reduction
Reduction Through
Through Formation
Formation
Flight
Flight
TRL LEVEL 7
TRL LEVEL 3
Theory 50% Reduction
in Induced Drag
Experimental Early FF-18 Data
Shows 1010-15%
Total Drag Loss
''C
CDDii == 35%
35%
''C
CDD == 10
10 –– 15%
15%
''w
wff == 10
10 –– 15%
15%
For a transcontinental route,
per trailing aircraft per year
Safety
Reliability
Feasibility
Ready for
Commercial
Application
''$$ == 0.5M
0.5M
''CO
CO22 == 10M
10M lbs
lbs
''NO
NOxx == 0.1M
0.1M lbs
lbs
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 6
Autonomous Formation Flight
Partners and Responsibilities
NASA DFRC
- Overall Project Management
- Flight Safety and Mission Assurance
- GN&C Design and Analysis
- Verification and Validation Testing
- Flight Vehicle Integration
- Flight Test Operations
Project Has NASA RevCon Status
And Is Reported At The
Congressional Sub-Committee Level
UCLA
Theoretical Research
- GN&C Design Methodologies
- GPS Algorithm Development
- Advanced System Concepts
The Boeing Company
• Operational Concept
• GN&C Design and Analysis
• Aerodynamic Models and Simulations
• Formation Flight Information System (FFIS) (Integrated GPS & IMU).
• Formation Flight Computer System (FFCS).
• Formation Flight Control System Software.
• Integration with F-18 Flight Control Computer (PSFCC) Systems.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 7
Revolutionary Technologies
Develop Three Key
Technologies:
•
Relative
Navigation
1st Close Coupled
Differential Carrier
Phase GPS-IMU
capable of 10 cm
relative accuracy.
Vortex Induced
Drag Reduction
Formation
Control
•
The 1st operational
formation drag reduction
tests under complete
auto-pilot control.
•
The 1st coordinated formation
flight of an auto-pilot controlled
aircraft to within sub-meter
accuracy.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 8
AFF Development Roadmap
2000
Phase 0
2001
2003
Funding: $13M Over 4 Years
AFF
AFF Station
Station
Keeping
Keeping
12/00
2002
Phase 1
Aero-Vortex
Aero-Vortex Mapping,
Mapping,
Bandwidth
Assessment,
Bandwidth Assessment,
Differential
Differential Carrier
Carrier Phase
Phase
GPS/INS
Demo
GPS/INS Demo
06/01
AFF
AFF Drag
Drag
Reduction
Reduction Flight
Flight
Demonstrate Functionality
of the Differential Carrier Phase
GPS/INS Hardware & A/C Telemetry.
Tests Demonstrate Complete Functionality
of the AFF System, Flight Control Avionics,
Differential Carrier Phase
GPS/INS Hardware and Aircraft Telemetry
07/02
Phase 2
AFF
AFF Transport
Transport
Flt
Flt Conditions
Conditions
&
& Ops
Ops Demo
Demo
AFF
AFF Optimal
Optimal
Performance
Performance
Demo
Demo
Autonomous
Autonomous
Aerial
Aerial
Refueling
Refueling
04/03
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 9
Program Approach
Create the Autonomous Formation Flight Project (AFF) Using
two NASA F/A-18 airplanes
• Phase 0 - Demonstrate Autonomous Station-Keeping
– Fall of 2000
• Phase 1 Risk Reduction - Map the Vortex Effects
– Fall of 2001
• Phase 1 - Autonomous Formation Flight
– Incomplete
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 10
Lift and Drag Force Basics
L
Resultant
Aerodynamic Force:
§¨ L2 D 2 ·¸
©
¹
D
D
Flight Path
•
V
Aerodynamic forces on an aircraft
– Drag is parallel to flight path
– Lift is perpendicular to flight path
Figure not to scale
– Lift is an order of magnitude greater than drag
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 11
Vortex Influence on Lift and Drag
'D
L’
Rotation Effect
of upwash (W)
L
Resultant
Aerodynamic Force:
§¨ L2 D 2 ·¸
©
¹
D’
'L
D
D
V’
W
Flight Path
V
•
'D
Figure not to scale
Basic theory states drag reduction, 'D, is caused by the rotation of the
original lift vector due to the upwash effect of the vortex
– The associated lift increase is very small because D<<L
– Only the induced drag is affected by vortex, 'D = sin('D) L
The most common theory on Formation Flight states that “drag reduction” is actually obtained due to a rotation of the lift vector that occurs while a trailing aircraft is in the upwash field of the lead
aircraft. The figure above illustrates this concept showing how the baseline (non-formation flight) lift and drag values, L and D, rotate by the change in angle of attack, 'Į, due to the upwash effect
while in the vortex flowfield.
Because of traditional bookkeeping methodology, the actual lift and drag values are maintained relative to the vehicle’s global, rather than local, flight path during formation flight. The term, 'D, is
used to represent the drag change due to the rotation of the lift force from L to L’. The drag during formation flight, DFF, is obtained by:
DFF = D’ cos('Į) -'D where: 'D = sin('Į) L
In a similar manner the term 'L, is used to represent the lift change due to the rotation of the drag force from D to D’. The lift during formation flight, DFF, is obtained by:
LFF = L’ cos('Į) + 'L
where:'L = sin('Į) D
Because lift tends to be an order of magnitude greater than drag (L>>D), drag is influenced significantly more by the rotation effect than lift is. A considerable reduction in drag can be realized by a
small upwash angle, while an insignificant increase in lift occurs.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 12
F-18 Wing Vortices & Cross Flow Gradient
3-View, F/A-18E: Mach 0.85, AOA 3deg
Trailing Aircraft In This Wake Experience
An Asymmetric, Turbulent Flow Field
Contours of
Pressure Coeff (Cp)
0.00
-0.25
-0.50
CFD Results: Courtesy of Dave Stookesberry, Boeing STL.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 13
Vortex Influence on Induced Drag
Percent Induced drag change, M=0.56, 25,000 ft, 55 ft N2T
0%
0.4
4
Flight Test
-4
-4
12
-12
25
-25
0
-0.4
-0.5
-40
Larger “sweet spot”
0
Lateral Separation (Y), wingspans
0.5
Calculated induced drag change
obtained from flight data, with
similar results at ALL flight conditions!
Vertical Separation (Z), wingspans
Vertical Separation (Z), wingspans
Rapid Drag Increase
0.4
4 4 0% 0%
-4
-4
Theoretical
-12
-12
-25
-50 -25
-50
0
-0.4
--0.5
0.5
0
Lateral Separation (Y), wingspans
Predicted induced drag change using
generic horseshoe vortex model*
*Adapted from: Blake and Multhopp, AIAA-98-4343, August 1998
Hammer home that this is INDUCED drag!
The flight results also measure higher drag increases inboard than predicted, but this is also the region where data quality is worse because the points are more difficult to fly. Some of these points
were very unstable as the vortex seemed to impinge on the tail or other surfaces causing the trailing aircraft to continually wander from the target position. Higher trim drag effects could
also contribute to the large drag increases. The line of zero benefit is also located further outboard than predicted. These results indicate substantially higher sensitivity to lateral positioning
inboard of the sweet spot than predicted. Small changes in lateral positioning in this region can result in large changes in benefits (drag increase!). The overall vertical sensitivity
is less than predicted; the overall shape of the region of most benefit is more round than oval as predicted for a generic wing. Induced drag results are similar at all flight
conditions and separation distances:
The induced drag change measured at the transport flight condition (not presented) correlated very well to those obtained at the reference condition shown above in both shape and magnitude. This
is a significant result indicating an accurate model of induced drag change could potentially be used to model drag benefits at other conditions.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 14
F-18A Wake Vortex Characteristic Aero-Increments Vary
Greatly With Offset Distance 'Y Between A/C
Induced Pitching Moment (I 0)
Induced Yawing Moment (I 0)
0.05
0.012
'Z
0.010
0.008
0.006
0.004
'Cn
0.002
0.000
-0.002
-0.004
-0.006
100
200
300
400
500
600
700
800
900
+280
+240
+200
+160
+120
+080
+040
+020
+010
+005
+000
-005
-010
-020
-040
-080
-120
-160
-200
-240
-280
0.03
0.02
'Cm
0.01
0.00
-0.01
-0.02
-0.03
-0.04
-0.008
0
'Z
0.04
+280
+240
+200
+160
+120
+080
+040
+020
+010
+005
+000
-005
-010
-020
-040
-080
-120
-160
-200
-240
-280
0
1000
100
200
300
500
600
700
800
900
1000
'Y
'Y
Induced Rolling Moment (I 0)
Drag Reduction (I 0)
3.00
0.03
'Z
'Z
0.02
+280
+240
+200
+160
+120
+080
+040
+020
+010
+005
+000
-005
-010
-020
-040
-080
-120
-160
-200
-240
-280
0.00
-0.01
-0.02
-0.03
-0.04
Optimal, Min Drag Near
Point Where Wingtips Align
2.50
CD form / CD
0.01
'Cl
400
2.00
+280
+240
+200
+160
+120
+080
+040
+020
+010
+005
+000
-005
-010
-020
-040
-080
-120
-160
-200
-240
-280
1.50
1.00
0.50
-0.05
0.00
0
100
200
300
400
'Y
500
600
700
800
900
1000
0
100
200
300
400
500
600
700
800
900
1000
'Y
Linear Panel Method Results
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 15
AFF Research Aircraft
NASA 845 Systems Research Aircraft (SRA)
• Pre-Production TF-18A (2 Seater)
• Research Modifications
– Instrumentation/Telemetry System
– Independent Separation Measurement System
– Formation Flight Control System & Instrumentation
System
– Production Support Flight Control Computers
– Engines Modified with Flight Test Instrumentation
Package for Thrust Measurement
– Cockpit Highly Adaptable Research Monitor System
– HUD Video & Hot Microphone System
NASA 847
• Production F-18A (1 Seater)
• Research Modifications
– Instrumentation/Telemetry System
– Independent Separation Measurement System
– Formation Flight Control System & Instrumentation
System
– Production Support Flight Control Computers
– HUD Video & Hot Microphone System
Two NASA F-18 aircraft were used for this research. Both aircraft were equipped with instrumentation and telemetry systems as well as identical GPS receiver units. The
Systems Research Aircraft (SRA) was designated as the follower and outfitted with the formation autopilot, consisting of a research computer and specially modified flight control
computers. A NASA chase aircraft acted as the formation lead. A third NASA chase aircraft was occasionally used for photographic documentation of the experiment.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 16
NASA’s F-18s Are Uniquely
Modified Production Versions
•AFF Avionics Tied Into F-18
A/C Bus Directly
•Boom & Drogue Refueling
•Fully Instrumented
Engines, Inlets, & A/B
•F-18 A Production
Equipped Avionics,
Digital 4x FCS, GPS,
RLG-IMU.
•AFF 2 Mb/s 9GHz Inter-Ship LAN
•NASA-EAFB Flight Test Telemetry
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 17
AFF System H/W Couple The Aircraft
Through A Wireless LAN
ISMS
Pilot Interface
CPB
Independent
Safety System
Outer-Loop Guidance
and Control
FFCS
AMUX
Multiplex / Filter
PSFCC
Inner-Loop Control
Envelope Monitoring
Differential Carrier Phase GPS
& Inter-ship Communication
FFIS
Wireless LAN Connection (9 GHz, 2.1 MB/sec)
Trail Aircraft
Lead Aircraft
FFIS
CPB
FFCS
AMUX
PBD
ISMS
PSFCC
Cockpit Interface
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 18
AFF Guidance Overview
Two Guidance Approaches
Trajectory Tracking
Leader-Follower
-'Z
-'X
+'Y
Yf
Zf
•
•
•
Trajectories defined by great circle
path. IC = lead aircraft initial heading,
velocity and alt.
Position errors are calculated between
AC and prescribed trajectory.
Appropriate for small and large
formations with prescribed
maneuvering.
•
•
•
Xf
Reference frame defined by lead
aircraft’s current velocity vector.
Position errors are based on aircraft
relative position.
Appropriate for tracking arbitrary
maneuvering. Potential Application
To Aerial Refueling & Auto-Carrier
Landing Systems.
Autonomous Formation Flight Program
NAS4-00041 TO-104
Page 19