15 Automatic Ground Collision Avoidance

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 43

ACAT FRRP

Automatic Collision Avoidance Technology


Fighter Risk Reduction Program

Automatic Ground Collision Avoidance:


Research to the Real Thing

Copyright 2009 Lockheed Martin Corporation.


F-16C, 15 Jun 07
Night Spatial Disorientation

• F-16 IP with 1582 hrs


in the F-16C/D and
159 hrs combat.

• No attempt to eject.

AFRL-WS 06-0093
Auto GCAS Development History
Advanced Fighter Technology Integration (AFTI F-16)

Nuisance Criteria You Are Here


Developed

1985 1990 1995 2000 2005 2010


1984: Initial Development 2007 – 2008: Block 60 F-16
• Test safety aid for AMAS (Auto • Full envelope, all loadings
Maneuvering Attack System)
• Throttle backdrive, Autopilot hand-off
• Partial envelope 1991–1993: Follow-on Development
• Flat Earth • First integration in operational fighter
• Safety aid for CAS/BAI
• Partial envelope
1997–1998: Full Envelope AGCAS
• Digital terrain database (AFRL + Swedish Government)
• Led to DTS and PGCAS
incorporation in production F-16s  Jan 97 – Oct 97: Nuisance criteria testing.
Defined boundary below which recovery
should be initiated
 Jul 98 – Nov 98: Full Auto-GCAS.

The 80% Solution

Prevent 80% of F-16 CFIT


Nuisance Free to 200’ AGL
Cost of Not Fielding Auto GCAS
USAF Fighter ACAT preventable losses since Auto-GCAS Maturity
• 19 Mar 00 CFIT Fatality F-16 destroyed
• 28 Aug 00 CFIT Fatality F-16 destroyed
• 12 Jun 01 Spatial-D Fatality F-16 destroyed
• 06 Jul 01 G-LOC Fatality F-16 destroyed
• 17 Jul 01 Spatial-D 2xFatal F-16 destroyed
• 09 Sep 02 Spatial-D Fatality F-16 destroyed
• 13 Nov 02 Spatial-D Fatality F-16 destroyed
• 09 Sep 03 CFIT Ejection F-16 destroyed
• 14 Sep 03 CFIT Ejection F-16 destroyed
• 05 Apr 06 G-LOC Ejection F-16 destroyed
• 27 Nov 06 CFIT Fatality F-16 destroyed
• 15 Jun 07 Spatial-D Fatality F-16 destroyed
• 15 Jan 08 Spatial-D Ejection F-16 destroyed
• 14 Mar 08 G-LOC Fatality F-16 Destroyed
• 25 Mar 09 CFIT Fatality F-22 Destroyed
ACAT FRRP Goals
• Reduce Fleet wide Integration Costs/Risks
– Make it better
• F-16 demands more protection, less nuisances, no Pilot
interaction
• No OFF, No disengage…..“PERFECT” system?

– Common architecture/modules for F-35, F-22, F-16


– Common Requirements, Design & Evaluation guides

DO NO HARM DO NOT INTERFERE AVOID COLLISIONS


1998 Architecture
(80% Solution)

State
Info
INS ARM FLCS
TRN Collision
Position Detect
Vector

DTED SCAN
Controls
Displays

6 arc Sec
~600’
Auto-CA Architecture
(98% Solution)

TPA FLCS
EGI
Collision
Detect

SRTM SCAN
Controls
Displays

6 Arc Sec
DTED Accuracy Error
(Large Error)
DTED Accuracy Error
(Large Error)
Nuisance Criteria
Error Budget

Nuisance Threshold

Mission
Command Guidance
Nuisance Criteria

“Error Budget”
NUISANCE CRITERIA TESTING
Error Budget
1) Keep errors as small as possible - They add up quick
2) Assess the errors
• Non Issue Errors
• Small errors can be “padded” to guarantee protection and
retain nuisance free operation - Avoid padding pads!
(sTotal)2= (sDTED)2 + (sTPA)2 + (sNAV)2
• Large errors require a tradeoff
• Nuisance Free Operation (Do not interfere)
• Provide Protection (Avoid collisions)

ACAT: Pilot Controlled Protection Level - “HI” , “LO” modes


HI = Guaranteed Protection
(nuisance free for all elements except Low Angle Strafe & Terrain Masking <500’)
LO = Guaranteed Nuisance Free (200’ mission)
NAV Improvements
1998 “80% solution”
Navigation Element
Sim
CTUS/H

3 Arc S

Pilot View Navigation Solution

Large Error > Error Budget


EGI Performance
Cal City Tank
Calibration Point

EGI Derived Left Wing Tip


EGI Derived Right Wing Tip

Small Error < Error Budget


TPA Improvements
• Predict recovery to very high fidelity
– Run ~180 times faster than real time
– No Auto-throttle: Do not know what pilot will do with throttle
• Complicates high speed recovery prediction
• An assumption must be made for climbing recoveries
• No PLA on MUX so an estimate must be made about current PLA
– Large store loading matrix across significant gross weight band
– CAT I and III loadings
– Lateral asymmetries up to 25,000 ft-lbs

• Latency issues of aircraft state data effects accuracy of


prediction (Roll Axis of fighter aircraft)
“Arrested Roll Angle” significant determinant of Altitude Lost

Nuisance criteria extremely useful to “Tune”


Clean GLOC (AB)
Utility Loading Needles
DTED Errors > Error Budget
CANNOT be overcome with creative design

Pre SRTM Data Terrain Survey

DTED Management
Is an Ongoing Requirement
SRTM
3 4
Terrain Survey Results
5

1 2

SRTM
Max Pre SRTM
400 400

Avg
200 200

0 0
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Min

-200 -200

-400 -400

-600 -600

-800
-800

-1000
-1000
Needles DTED
1” SRTM v 1/3” NED
Resulting Difference
Elevation Shift is from 0ft to 436 feet. SRTMF DTED was only shifted up not down.
DTED Conclusions
• Pre-SRTM DTED
– Unsuitable for Nuisance free operation below 500’/200’
(Rough/Smooth Terrain)
– Get SRTM data from NGA

• SRTM Data
– Capable of supporting 200’/100’ Nuisance free operations
(rough/smooth terrain)
– Enhancements required to provide full protection over select features
in a low level mode. DTED Management!

• SRTM Inaccuracy
– Average Error (Small)– Add Pad
– Errors on Peaks (Large but isolated) – Tradeoff in favor of protection in
HI, in favor of nuisance free in LO
– Foliage Study Ongoing
DTED Resolution
DTED Resolution
DTED Resolution Induces Errors
Result: Nuisance fly-ups for terrain that is
actually well below aircraft

6 Arc Second
3 Arc Sec – No Nuisance

3 Arc Second
1 Arc Second
DTED Resolution Drives Scan Width
Result: Scans non factor terrain
First DTED Post scanned
If Scan pattern optimized

Non factor terrain scanned


Nuisance Flyup Generated

Aircraft Position
Encoded DTED gives higher resolution
where needed & smaller file size

r
Origin
o
Radius

Very Accurate, Very Small Size Poor Accuracy, Very Large Size
“Encoded” CONUS
328 Kbytes

31
Coupler Challenge
• Large store matrix
• No “active g-limiter”
• CAT I vs. CAT III
• Large lateral asymmetries

• This was not considered a technical challenge!


• Planned to use existing coupler from 1998 demo
Original Coupler
Coupler Design Improvements
• Longitudinal:
Zero g unload vice -1g push
When at high g, Unload prior to beginning roll

• Lateral:
Roll-Through logic improved to prevent dwells inverted
G-limiter as a function of roll rate
Transonic Improvements
External wing tank rolling g limits >.95M with fuel

• Reduced altitude lost


• Significantly better ride quality
• Significant improvement in Over-g protection
Improvement Areas

• Improvements in NAV solution (EGI)


• Improvements in TPA (latency and rolling)
• Improvements in DTED (3 arc second, SRTM+)
• Improvements in Scan (3 arc second scan)
• Improvements in Coupler (Over-g Protection)
Lessons Learned

• Develop, evaluate, refine Auto-maneuver first


– Trajectory predictions, scan patterns depend on this
• Don’t assume that something tested previously isn’t broke
• Garbage in, Garbage out
– Cannot avoid a hill you do not know exists
– DTED is King
– Do not accept Status Quo – DTED is not “off limits”
Questions???
Improvement Areas

• Improvements in NAV solution


• Improvements in TPA
• Improvements in DTED
• Improvements in Scan
• Improvements in Coupler
20,000 ft-lbs Asymmetry
Comparison of Coupler Designs
5

4 New design:
Nose continues to
drop while inverted
3

More than make up


2 for it with a faster
Nz onset of positive g
New
1
Average:
21.4 feet less alt loss
0 Old

-1
0 1 2 3 4 5 6 7 8 9 10

Time (sec)
“Lateral” Escape

You might also like