The Wide Area Augmentation
System (WAAS)
Todd Walter
Stanford University
http://waas.stanford.edu
Conclusions
WAAS is used to provide aircraft
navigation from enroute through
vertically guided approach
Integrity was and is the key challenge
Important to understand what can go
wrong and how to protect users
Careful analysis of feasible threats
2
New civil frequencies and additional
constellations will further improve
performance
Outline
The Wide-Area Augmentation System
Integrity Analyses
Comparison with Terrestrial
Navigational Aids
Future Directions
3
WAAS
4
WAAS Architecture
Wide Area Augmentation System (WAAS) – Program Status
38 Reference
3 Master
4 Ground
Stations
Stations
Earth Stations
2 Geostationary
2 Operational
Satellite Links
Control Centers
Courtesy:
Federal Aviation
Administration
5
Geostationary Satellites (GEO)
PanAmSat
133W
•
Telesat
107W
Provides Dual Coverage Over United States
Wide Area Augmentation System (WAAS) – Program Status
Courtesy:
Federal Aviation
Administration
6
Wide Area Augmentation System (WAAS) – Program Status
Federal Aviation
Administration
7
Error Sources
Satellite errors
Ephemeris
Clock
Signal
Propagation errors
Ionosphere
Troposphere
Local Errors
8
Multipath
Receiver Noise
Master Station Schematic
9
Complicated Schematic
10
GPS Performance (Usually)
On a good day, the
red circle encloses 95%
of the GPS position fixes.
11
Major GPS Faults About Twice a Year
Example: Ephemeris Failure on April 10, 2007
On a bad day, the
GPS errors can be
much worse.
WAAS & GBAS
eliminate these
large errors.
12
Integrity Approach
Aviation integrity operates on a guilty
until proven innocent principle
Error bound is the maximum possible
value given the measurements
This is unlike conventional systems
that describe the most likely errors
Protection level is a 99.99999%
bound on worst reasonable conditions
13
Very different from 95% achieved
accuracy
Failure of Thin Shell Model
Quiet Day
14
Disturbed Day
Undersampled Condition
Courtesy:
Seebany
Datta-Barua
15
11/20/2003
21:00:00 GMT
16
Localizer Performance Vertical (LPV) Coverage
Wide Area Augmentation System (WAAS) – Program Status
Courtesy:
Federal Aviation
Administration
17
WAAS RNP 0.3 Current Coverage
Wide Area Augmentation System (WAAS) – Program Status
Courtesy:
Federal Aviation
Administration
18
WAAS LPV and LPV-200 Vertical Position
Error Distributions July 2003 to June 2006
Courtesy:
FAA
Technical
Center
3 years
20 WRSs
1 Hz data
19
Navigational Aids
Instrument Landing System (ILS)
Glideslope antenna for vertical
Localizer for horizontal
20
ILS Installations: Each Runway End
Requires At Least Two Transmitters
1318 ILS’s nationwide
21
No GPS Equipment Required at Airport
50 Pieces of WAAS Equipment Serve the
Continent
As of November, 2009
• 1820 WAAS-based LPV’s
• ~1000 for non ILS runways
22
Localizer Approaches at
Moffett Field
Courtesy:
Sharon Houck
23
Utility of Protected Accuracy from WAAS
• Localizer performance with
vertical guidance (LPV)
• Safer than lateral nav.
(non-precision approach)
• Same decision ht. as Cat I
• GBAS for Cat. II & III
WAAS (& GBAS) tunnels:
• Do not flare like ILS
• Do not have beam bends
• Are programmable
• Are adaptable
24
Current WAAS Performance
25
Future L1/L5 Performance
26
L1-only Threats
 Reference station multipath, noise, cycle slips, and
other receiver errors – WRE bias, CNMP
 Satellite clock/ephemeris – UDRE/MT28
 Erroneous ionospheric delay estimates - GIVE
 Erroneous WRE clock estimates, interfrequency
bias estimates – RDM
 L1 code/carrier incoherence – CCC
 L1 signal deformation – SQM
 Multiple convolved threats – UPM, convolution
analysis
 Antenna biases, GEO biases – specific analyses
27
L1/L5 Threats
 Reference station multipath, noise, cycle slips, and
other receiver errors – WRE bias, CNMP –
unchanged
 Satellite clock/ephemeris – UDRE/MT28 – need to
remove L1/L2 bias from Fast Correction, could
remove uncertainty from UDRE
 Erroneous ionospheric delay estimates - GIVE –
uneeded
 Erroneous WRE clock estimates, interfrequency
bias estimates – RDM – IFBs uneeded, WRE
clocks may be handled by UDRE/UPM
28
L1/L5 Threats cont.
 L1/L5 code/carrier incoherence – CCC –
new/updated monitor, MERR reduced without
GIVE while threat is potentially increased
 L1/L5 signal deformation – SQM – new/updated
monitor, MERR reduced while threat is increased.
User space will need to be greatly restricted
 Multiple convolved threats – UPM, convolution
analysis – new/updated monitor specific to L1/L5
user
 Antenna biases, GEO biases – specific analyses –
new/updated analysis, effects are greater
29
CCC Threats
Iono-free ccc metric expected to have 2.6
times as much noise as L1-only version
Trip thresholds may need to be increased
Could lead to UDRE bumps or trips
MERR substantially reduced without GIVE
Need much reduced thresholds and test
L1 only monitor has margin thanks to GIVE
Need to collect data to see real test
Need L5 data, but L2C might be OK
High risk factor for not being able to match
L1 UDREs
30
Frequency Dependencies
L1
Only
Ionofree
L1 error L5 error
RSS
Sum of
Iono Clock /
absolutes error ephemeris
or tropo
1
1
1
1
1
2.6
3.5
0
1
0
f12
f12  f52
f52
f12  f52
 2.26
 1.26
PRL1ij  R i gl ij  Bi  I L1ij  T ji  b j  M L1ij   L1ij
 L1ij  R i gl ij  Bi  I L1ij  T ji  b j  N L1ij 1  mL1ij  nL1ij
PRL1ij  R i gl ij  Bi    I L1ij  T ji  b j  M L 5 ij   L 5 ij
31
 L 5 ij  R i gl ij  Bi    I L1ij  T ji  b j  N L 5 ij 1  mL 5 ij  nL 5 ij
SDM Threats
Bias threats on L1 have 2.26 times greater
influence on iono-free users
L5 biases also must be added
MERR substantially reduced without GIVE
Nominal and threat biases potentially 3.5
times those on L1 only
Bias terms in VPL potentially very beneficial to
iono-free user
Need to further restrict user space
32
L1 and L5 may need to be a very tight box
around monitor choice
Difficult to get international buy-in
Conclusions
WAAS is used to provide aircraft
navigation from enroute through
vertically guided approach
Integrity was and is the key challenge
Important to understand what can go
wrong and how to protect users
Careful analysis of feasible threats
33
New civil frequencies and additional
constellations will further improve
performance
Potential of L1/L5
GPS/Galileo Performance
34