Industry Weighs in on Error-Prone SVN-49

Industry Weighs in on Error-Prone SVN-49
June 15, 2010
By: Alan Cameron

http://www.gpsworld.com/gnss-system/gps-modernization/news/industry-weighs-=
error-prone-svn-49-10069?print=3D1

The public comment period on proposed mitigation options for GPS
satellite IIR-20M (SVN-49) ended May 28, and comments are viewable at
www.regulations.gov under RITA Docket 2010-0002. Among others, the
U.S. GPS Industry Council, NovAtel, Garmin, Septentrio, Raytheon,
Boeing Commercial Airplanes division, General Motors OnStar, the
European Commission, the MITRE Corporation, STMicroelectronics, the
German Space Operations Center, and Cessna Aircraft have all filed
comments expressing a preference for one option or another.

Unfortunately for the U.S. Air Force and the GPS Wing, no clear
consensus emerges. Indeed, differences of opinion naturally follow the
respective orientation of each company or organization toward their
customers=92 or members=92 specialized needs.

This article recaps salient commentary from several companies and
organizations, concentrating on their preferred option.

The potential mitigations are each designed to reduce the impact of
the unique nature =97 that is, errors =97 of the SVN 49 signal to a
portion of the user segment. The Air Force GPS Wing actively sought
public and industry input regarding concerns or adverse impacts that
may affect civil GPS user equipment.

For background on the SVN-49 situation, see Richard Langley=92s Expert
Advice column from August 2009. Very briefly, the pseudorange data
broadcast by the satellite contains larger than normal errors that
vary according to the elevation of the satellite above the horizon.

The mitigation techniques that have been proposed are:

Option 1. Set healthy with current 152-meter antenna phase center
(APC) and associated clock offsets.
Option 2. Set healthy with factory APC offset.
Option 3. Users switch to multipath-resistant receivers.
Option 4. Modify receiver software to use look-up table corrections.
Option 5. Increase user range accuracy (URA) index to a minimum value
of 3.
Option 6. Remove data modulation from L2 P(Y)-code, and
Option 7. Change L2C PRN code to a =93unique sequence.=94
(Options 6 and 7 are normally considered a pair; that is, they must be
jointly implemented to have the desired effect.)
Option 8. Change SVN-49 from PRN-01 to PRN-32.
Option 9. Use spare health code so future users could use SVN 49
despite unhealthy setting.
U.S. GPS Industry Council. The comments filed by the U.S. GPS Industry
Council (USGIC), available as a PDF file here, are the most detailed
and extensive across all the options. However, the stated preference
of the USGIC for Option 9 does not necessarily reflect agreement
across all sectors of industry. As the USGIC points out, =93Options 1
through 8 propose to designate SVN 49 as healthy using techniques that
enable mitigation for some user applications, but that are unable to
also mitigate adverse impacts to other users.=94

After discussing historical background and perspective, including
sections on differential, carrier phase, and non-compliant receivers,
and further consideration of each individual option, the USGIC
concludes that =93For about two decades, the installed user base has
experienced a GPS constellation of satellites set operationally
healthy with the resulting benefit that no users obtain erroneous
answers. Consequently, we find Option 9 to be the most appealing
option of the proposed mitigations presented. Manufacturers who choose
to do so could produce firmware that corrects for the SVN 49 signals
and enable use of satellites displaying the unique health code. This
implementation allows for the potential to derive some value from this
SV in the future, without the attendant liabilities inherent in the
other proposed mitigations (Options 1 through 8).

=93[USGIC] Recommendation: Either maintain SVN 49 as unhealthy, or put
Option 9 into effect for SVN-49 to sustain GPS availability as needed
in the operation of the constellation in the future. It is the only
option that fully protects all applications, and the only option that
ensures that no harm is done to differential users.=94

NovAtel. This high-precision manufacturer recommends adopting Option
5, increase URA to 3 or greater. =93The benefit of the additional
geometry outweigh[s] the positional errors that may be incurred in all
modes of operation. Being able to provide a position (with quantified
error uncertainty) is better then not being able to provide any
position at all.

=93L1C/A impacts: Our L1C/A users primarily operate in DGPS mode and
would ignore this value. Anomaly range error (<2m) is not large enough
to cause problems with our L1 only RTK. Dissimilar reference station
to receiver correlator types will see the anomaly.

=93L1C/A semi-codeless: Our RTK users ignore this value. Anomaly range
error is not large enough to cause problems with our L1/L2 RTK.
Anomaly carrier phase error (<3mm) is insignificant and cancels in the
processing.=94

NovAtel does not recommend Option 9, the USGIC=92s preference, stating
that =93The problem is not severe enough for us to write software to
give this SVN special treatment,=94 and citing Option 9=92s impact as
=93reduced geometry.=94

Garmin. The company with a strong presence in the consumer handheld
market, but also a longtime stake in general aviation, prefers Option
2, set healthy with a factory APC. =93It minimizes the impact to the
performance of receivers with orbital predictions.=94 Garmin also
expresses a predilection for Option 5, NovAtel=92s choice, but =93prefers
this is implemented along with mitigation #2.=94

Raytheon. The prime contractor for the development and maintenance of
the U.S. Wide Area Augmentation System (WAAS) focuses its comments on
user needs of this and other satellite-based augmentation system
(SBAS) equipment, for safety-critical applications such as precision-
approach aircraft landing.

=93Safety Critical SBAS systems like WAAS in North America, EGNOS in
Europe, MSAS in Japan and GAGAN in India are all adversely affected by
the anomalous signals on PRN 1. SBAS systems strive for sub meter
accuracy. If PRN 1 were set healthy without any modification to the
processing algorithms on the ground the satellite would cause frequent
alarms and other degraded performance. Some SBAS Service providers
have taken steps to avoid this behavior by removing PRN#1 from the
system. In addition to SBAS, Raytheon has other high integrity local
area differential GPS systems under development that may be adversely
affected by PRN 1. . . . A long term solution for PRN 1 is required to
ensure that a replacement satellite in the 2020 timeframe does not
adversely affect these systems.=94

Raytheon does not state an explicit preference among the options; it
seems to find downsides for each of them. Indeed, one of Raytheon=92s
biggest SBAS customers, the Federal Aviation Administration, has
indicated that it will not use SVN-49 at all. Raytheon finds no pros,
but significant cons, for Options 1, 3, 6, and 7. For Option 2, set
healthy with factory APC, Raytheon grudgingly admits =93SBAS system may
be able to utilize PRN 1 with a large uncertainty term (large UDRE),=94
but foresees =93significant modifications to the SBAS ground systems=94
that will take on the order of a year to develop, field, and test.

The company appears to lean toward Option 4, user equipment software
updates, finding in its favor a =93possibility of a long term solution
for dual frequency SBAS users in the 2018 time frame. (Dual frequency
user specs are not yet developed.)=94 Perhaps the strongest factor in
4=92s favor for Raytheon, SBASs, and their users is that they =93will not
make changes to accommodate the satellite in the near term . . .
Changes will be made to the ground system to disable the satellite
avoiding changes to user avionics.=94

Raytheon finds two further options less objectionable than the others:
Option 5 (=93Significant modifications to the SBAS ground systems if PRN
1 is to be used for precision approach operations=94) and Option 9
(=93Significant modifications to the SBAS ground systems if PRN 1 is to
be used for precision approach operations, however no changes are
required for user receivers.=94)

The U.S. Air Force Space and Missile Systems Center (SMSC) Small
Diameter Bomb Increment I unit, likely to swing a heavy cat in the
discussion room, rather tersely nods in the direction of Option 1, the
152-meter APC. Identical comments were also posted by Joint Direct
Attack Munition (JDAM), a U.S. joint armed forces organization
developed by the Air Force and the Navy. JDAM is a guidance kit that
converts unguided gravity bombs, or dumb bombs, into all-weather smart
munitions guided by an integrated inertial guidance system and a GPS
receiver.

Speaking as one, the SMSC and JDAM state, regarding Option 1, =93We
believe this will work without software change.=94 They find significant
L2 impact from and do not like either Option 2 (=93will cause errors=94)
or 3 (=93will not work in jamming=94). In their view, Option 9 requires
further study, a minimum of two years, as its impacts include =93Change
in GPS software and weapon OFP [Operational Flight Program].=94

Septentrio, a manufacturer of high-precision GPS receivers that also
has a strong presence in the Galileo effort, says that Option 5
=93should be done, but it is not a solution on its own.=94 The company
prefers Option 9: =93It makes sure that existing units deployed in the
field are not affected. Only updated receivers will be able to use
SVN-49. It should be very clear how SVN-49 can be identified until its
end of life. Updated receivers can use SVN-49 after correction of the
elevation angle dependent error using (receiver-specific) look-up
tables.

ST-Microelectrionics, which has a significant customer base within the
international automobile industry as well as other segments, prefers
Option 2, set healthy with factory APC, but gives no further
reasoning. It emphatically refutes Option 1, the 152-meter APC choice.
ST-Micro gives a =93yes=94 vote to Option 3, a mere =93ok=94 to 4 and 5,
expresses =93no concern=94 with 6 and 7, is =93happy=94 with 8, but votes
definitely no on 9: =93If any bit is set in the legacy ephemeris health
or legacy health packet, we will not use the satellite. Converse=85
healthy but some bit set in a new (not legacy) message would be ok.=94

The European Commission, speaking from the point of view of its SBAS,
the European Geostationary Navigation Overlay Service (EGNOS), finds
little impact from several options but pronounced adverse consequences
from others, including the USGIC=92s predilect Number 9.

=93The implementation of Action 8 is not recommended since it may (for
some period of time) impact the performances, without benefit in
counterpart for EGNOS Users.

=93The implementation of Actions 3, 4 and 9 may require an update of
user standards. Moreover it is unclear how to deal with possible
existing receivers sensitive to faulty/degraded PRN signal. In
addition, even implementing these actions at user level there are some
risks on System behaviour with respect to SVN-49 signal, which may
prevent its use. In conclusion, the implementation of Actions 3, 4 and
9 on the EGNOS user community (in case SVN-49 is eventually augmented
by EGNOS) will have major consequences on the existing EGNOS standards
and users with some risks/uncertain results.

=93The other mitigation actions look rather neutral for EGNOS so long as
SVN-49 is excluded from the EGNOS satellite mask.=94

regulations.govunder RITA Docket 2010-0002. Among others, the

This satellite is 'plain and simple' defective. After a year of study,
no one can agree upon a course of action to salvage it. Trying to use
it in a healthy constellation will cause confusion. SVN49 should be
relegated to residual status. It's only claim to fame is reservation
of the L5 signal frequency but I suppose it can still look out for
nuclear detonations. In those respects, it wasn't a total waste ..
that spin makes me dizzy.

--- CHAS