INNOVATION: Precise Point Positioning Apr 2009

The Business and Technology of Global Navigation and Positioning www.gpsworld.com

THE SYSTEM

Galileo Lite;
GLONASS
Tuned; GPS L5
Set to Go
THE BUSINESS

Security for
Military,
Mass Market
EXPERT ADVICE

IMES Evaluated
SYSTEM DESIGN

SBAS Interferes
BUSINESS OUTLOOK

How the Stimulus

Plan Benefits
Our Industry

VOLUME 20 | NUMBER 4
 SOLUTIONS AND PRODUCTS FOR A
NEW ERA IN SATELLITE NAVIGATION
... improved GLONASS ...

Modernized GPS ... ... GIOVE-B in orbit!

𰀩𰁎𰁃𰁒𰁅𰁁𰁓𰁅𰀀𰁙𰁏𰁕𰁒𰀀𰁐𰁒𰁏𰁄𰁕𰁃𰁔𰁉𰁖𰁉𰁔𰁙𰀀𰁗𰁉𰁔𰁈𰀀 𰀳𰁁𰁔𰁉𰁓𰁆𰁙𰀀𰁙𰁏𰁕𰁒𰀀𰁒𰁅𰁓𰁅𰁁𰁒𰁃𰁈𰀀𰁐𰁒𰁏𰁊𰁅𰁃𰁔𰀀𰁎𰁅𰁅𰁄𰁓𰀀𰁗𰁉𰁔𰁈𰀀
𰀧𰀰𰀳𰀀𰀬𰀑𰀌𰀀𰀬𰀒𰀌𰀀𰀬𰀕𰀀𰁃𰁏𰁍𰁍𰁅𰁒𰁃𰁉𰁁𰁌𰀀𰁐𰁒𰁏𰁄𰁕𰁃𰁔𰁓𰀁 𰁃𰁕𰁓𰁔𰁏𰁍𰁉𰁓𰁁𰁂𰁌𰁅𰀀𰀧𰁁𰁌𰁉𰁌𰁅𰁏𰀀𰀥𰀑𰀌𰀀𰀥𰀕𰀌𰀀𰀥𰀖𰀀𰁓𰁏𰁌𰁕𰁔𰁉𰁏𰁎𰁓𰀁

NavX®-NCS – RF Navigation
Constellation Simulator
Now Supports Galileo CBOC
and Japanese QZSS Signals
𰀴𰁈𰁅𰀀𰁬𰁒𰁓𰁔𰀀𰀧𰀰𰀳𰀏𰀧𰁁𰁌𰁉𰁌𰁅𰁏𰀀𰀲𰀦𰀀𰁃𰁏𰁎𰁓𰁔𰁅𰁌𰁌𰁁𰁔𰁉𰁏𰁎𰀀𰁓𰁉𰁍𰁕𰁌𰁁𰁔𰁏𰁒𰀌𰀀
𰁓𰁕𰁐𰁐𰁏𰁒𰁔𰁉𰁎𰁇𰀀𰁁𰁌𰁌𰀀𰁆𰁒𰁅𰁑𰁕𰁅𰁎𰁃𰁉𰁅𰁓𰀀𰁉𰁎𰀀𰁏𰁎𰁅𰀀𰁂𰁏𰁘𰀌𰀀𰁒𰁅𰁄𰁅𰁬𰁎𰁅𰁓𰀀𰁔𰁈𰁅𰀀
𰁆𰁒𰁏𰁎𰁔𰁉𰁅𰁒𰀀𰁏𰁆𰀀𰀲𰀦𰀀𰁃𰁏𰁎𰁓𰁔𰁅𰁌𰁌𰁁𰁔𰁉𰁏𰁎𰀀𰁓𰁉𰁍𰁕𰁌𰁁𰁔𰁏𰁒𰀀𰁔𰁅𰁃𰁈𰁎𰁏𰁌𰁏𰁇𰁙𰀎

NavX®-NCS RF Constellation Simulator

𰁳𰀀𰀀𰀧𰀰𰀳𰀀𰀬𰀑𰀌𰀀𰀬𰀒𰀌𰀀𰀬𰀕
𰁳𰀀𰀀𰀧𰀡𰀬𰀩𰀬𰀥𰀯𰀀𰀥𰀑𰀌𰀀𰀥𰀕𰀌𰀀𰀥𰀖
𰁳𰀀𰀀𰀧𰀬𰀯𰀮𰀡𰀳𰀳𰀀𰀬𰀑
Test Solutions
𰁳𰀀 𰀱𰀺𰀳𰀳𰀀𰀬𰀑
Receiver 𰁳𰀀𰀀𰀵𰁐𰀀𰁔𰁏𰀀𰀑𰀐𰀘𰀀𰁃𰁈𰁁𰁎𰁎𰁅𰁌𰁓𰀀𰁉𰁎𰀀𰁔𰁈𰁅𰀀𰁐𰁒𰁏𰁆𰁅𰁓𰁓𰁉𰁏𰁎𰁁𰁌𰀀𰁖𰁅𰁒𰁓𰁉𰁏𰁎
𰁳𰀀𰀀𰀵𰁐𰁇𰁒𰁁𰁄𰁅𰀀𰁃𰁁𰁐𰁁𰁂𰁉𰁌𰁉𰁔𰁙𰀀𰁂𰁙𰀀𰁁𰁄𰁄𰁉𰁎𰁇𰀀𰁃𰁈𰁁𰁎𰁎𰁅𰁌𰀀𰁍𰁏𰁄𰁕𰁌𰁅𰁓
Software Tools 𰁳𰀀𰀀𰀀𰀦𰁒𰁅𰁅𰀀𰁓𰁅𰁌𰁅𰁃𰁔𰁉𰁏𰁎𰀀𰁏𰁆𰀀𰁆𰁒𰁅𰁑𰁕𰁅𰁎𰁃𰁙𰀀𰁁𰁎𰁄𰀀𰁓𰁙𰁓𰁔𰁅𰁍

***Visit Us at Booth 10, ENC-GNSS 2009, Naples, Italy***

www.ifen.com
sales@ifen.com
 APRIL 2009
VOL. 20, NUMBER 4
gpsworld.com

THE SYSTEM 10 » COVER STORY

Big Four Race Called from Munich;
DAGR Extended; Jamming
Trial Confirms Vulnerabilities;
Briefly Noted

THE BUSINESS 14
Security Paramount, Strengthened
by Location; Broadcom, SiRF Top
ABI’s GPS Chip Vendor Ranking;
ITT Awards Symmetricom Timing
Contract for GPS III; more

BUSINESS
OUTLOOK 27
How the Stimulus Plan
Benefits Our Industry
How will the American Recovery
and Reinvestment Act of 2009
affect our industry?
By Stephen Colwell

SBAS Interference 37
OPINIONS & Worst-Case Scenario
Some applications using data collected by fixed- or low-dynamic
DEPARTMENTS WAAS or other SBAS receivers may encounter a tracking error due
to C/A code cross-correlation.
By L. Lestarquit, J-L. Issler, O. Nouvel, M. Sihrener, G. Lamontagne,
Out in Front 4 J.C. Guay, R. Landry, O. Julien, C. Macabiau, and M. Nouvel-Malicorne
Talking Heads
By Alan Cameron

Expert Advice 6
Indoor Message System Evaluated
Precise Point Positioning 43
By Börje Forssell
A Powerful Technique with a Promising Future
The IMES indoor messaging
Is there a viable alternative to real-time kinematic (RTK) positioning?
system aims to provide seamless
In this month’s column, we take a look at the technique of precise
positioning at the room level.
point positioning (PPP). How close is PPP to prime time?
By Sunil Bisnath and Yang Gao
Marketplace Classifieds 49

www.gpsworld.com April 2009 | GPS World 3

OUT IN FRONT
Talking Heads
W
e now feel ourselves dragged
into the second decade of
the 21st century — some of
us kicking and screaming. An edict has
arrived from on high: explore the use of
new media to further our mission and
increase our numbers, whether they
count revenue, pages, or web traffic.
Pick a number. Any number. Now
increase it.
Well and good. A former marketing
communications account exec, I’m all
for exploring the possible benefits of new
media, and the new ways of thinking
that they may instill. Exploring, talking
with those who have gone before, and
then evaluating the overall findings.
What follows does not constitute
only and mere navel-gazing. Replace my
enterprise with yours as you read. Would
your company or organization, or the
Would you get a reasonable ROI
of your time in Twitter, Linked-In,
IM, Facebook, YouTube?
GNSS industry as a whole, get a reason-
able return on investment of time spent
in Twitter, Linked-In, instant messaging
(IM) and other chat-forms, Facebook,
MySpace, YouTube?
Explore. My admittedly minimal
casting about in Linked-In and Facebook
suggests that Linked-In suits admirably
for job-finding and consultancy-fishing,
not meaningful lead- or traffic-genera-
tion. There isn’t enough room or context
for content, because it was designed for
the former two goals, not the latter two.
Facebook — you can go by its name.
Mostly about faces, and the superfici-
ality that accompanies appearances.
Can be fun, pass the time, keep up or
attempt to strike up with acquaintances
old and new. But not every human
pursuit naturally, eventually, or auto-
matically becomes a business pursuit
if practiced long enough. Facebook is
Letters to Editor invited: email to eic@gpsworld.com.
4 GPS World | April 2009
much akin to that relic of bygone
decades, the cocktail party.
Talk. Second method, I happened
across a fellow I know, who coinciden-
tally authored two cover stories in this
magazine and is otherwise active in the
GNSS community. Do you find social
media any use in your professional life,
industry, community? I asked.
“I generally don’t,” he wrote back,
“have any expectations for Facebook as a
professional tool. Some ‘grown-up’
versions of Facebook such as Linked-In
and Plaxo are going for that market. Even
those, I’ve been building a professional
network, but find it of little real use. If I
want to e-mail somebody, I e-mail them.
I don’t need a web-based thing for that.
“My corporate communications
policy limits the amount I can just blog
about work-related stuff. I couldn’t post
pictures or videos about my work if I
wanted to.
“Finally, I do this Facebook thing to
get away from the geek/nerd community
that I’m immersed in all day. If I needed
more outlets for that, there are plenty of
web resources that would be better.
“[Another GPSW author] related
an interesting observation when I
first started using Facebook. To him
it was disturbing when professional
acquaintances started asking to be
friends on Facebook because it was like
‘worlds colliding’.”
Evaluate. I’ll keep an open mind
but not invest more time at present.
Two years ago, I extolled the virtues
of blogging to deepen contact with
customers and other influentials in your
sphere (“Dagbloggit,” February 2007).
I stand by that; but so far, it doesn’t
extend to FB.
New Media. For now, I’m going with
a new medium that’s exactly about as old
as I am. Check out our website for videos.
Not Youtube, exactly. Short interviews
with leaders of the GNSS community.
Starting with Dr. Brad Parkinson and
some key Galileo figures. View on.
www.gpsworld.com
EDITORIAL
Editor in Chief Alan Cameron | acameron@questex.com
Managing Editor Tracy Cozzens | tcozzens@questex.com
Art Director Annie Leung | aleung@questex.com
EDITORIAL OFFICES
201 Sandpointe Avenue, Suite 500,
Santa Ana, CA 92707-8716 USA
714-338-6700 | Fax 714-338-6717
www.gpsworld.com | info@gpsworld.com
CONTRIBUTING EDITORS
Innovation Richard Langley | lang@unb.ca
LBS Kevin Dennehy | kdennehy@questex.com
Mass Market OEM Stephen Colwell | scolwell@questex.com
Military/Government Don Jewell | djewell@questex.com
Precision OEM Rob Lorimer | rlorimer@questex.com
Survey Eric Gakstatter | egakstatter@questex.com
Transport/Avionics Bill Thompson | bthompson@questex.com
Utilities/Communications Janice Partyka | jpartyka@questex.com
ADVERTISING
Publisher Kristina Panter
kpanter@questex.com | 714-338-6733
Eastern Region/International Account Manager Michael Eager
meager@questex.com | 714-338-6738
Western Regional Account Manager Jeannie Hortness
jhortness@questex.com | 714-338-6732
Classified Account Executive Candi Eager
ceager@questex.com | 714-338-6708
Marketing Coordinator Sarah Joy Obaña
sobana@questex.com | 714-338-6763
Sales Administrator Marisa Strange
mstrange@questex.com | 714-338-6766
ADVERTISING OFFICES
201 Sandpointe Avenue, Suite 500, Santa Ana, CA 92707-8716 USA
714-338-6700 | Fax 714-338-6717
PUBLISHING SERVICES
Production Manager Anna Abbott | aabbott@questex.com
Audience Development Manager Deborah Gullian |
dgullian@questex.com
PRODUCTION OFFICE 306 West Michigan St., Suite 200,
Duluth, MN 55802, USA; 218-279-8840
CIRCULATION/SUBSCRIBER SERVICES
gpsworld@halldata.com | USA: 866-344-1315 (toll free)
International: 1-847-763-9594
QUESTEX MEDIA GROUP, INC.
President & CEO: Kerry C. Gumas
EVP & CFO: Tom Caridi
EVP: Tony D’Avino
VP, Human Resources: Diane M. Evans
QUESTEX WORLDWIDE HEADQUARTERS
275 Grove Street, Newton, MA 02466, USA
617-219-8300 | Toll-Free 888-552-4346 | Fax 617-219-8310
MANUSCRIPTS: GPS World welcomes unsolicited articles but cannot be held responsible
for their safekeeping or return. Send to: 201 Sandpointe Avenue, Suite 500, Santa Ana,
CA 92707-8716 USA. Every precaution is taken to ensure accuracy, but publishers cannot
accept responsibility for the accuracy of information supplied herein or for any opinion
expressed. REPRINTS: Reprints of all articles are available (500 minimum). Contact
800-290-5460, ext. 100, e-mail GPSWorld@reprintbuyer.com. DIRECT MAIL LIST
RENTAL: Ilene Schwartz, Kroll Direct, 216-371-1667, fax 216-371-1669 e-mail Ilene@
krolldirect.com. SUBSCRIBER SERVICES: To subscribe, change your address, and all
other services, e-mail gpsworld@halldata.com or call 866-344-1315 (1-847-763-9594 outside
the U.S.). PERMISSIONS: Contact 800-494-9051 ext. 100 or questexpermissions@
reprintbuyer.com. INTERNATIONAL LICENSING: Contact e-mail info@gpsworld.com.
ACCOUNTING OFFICE and OFFICE OF PUBLICATION: 306 West Michigan
St., Ste 200, Duluth, MN 55802, USA.
GPS WORLD does not verify any claims or other information appearing in any of the
advertisements contained in the publication and cannot take any responsibility for any losses
or other damages incurred by readers in reliance on such content.
Published monthly
www.gpsworld.com
Dream Big...
Think Small.

Bring your big ideas down to size with our

space-saving semiconductor packaging
that makes your board designs possible.
Our new DDR2 SDRAM packs 1GB into
a SnPb ball attach BGA package less
than 1/2 in2.

WEDC lean multi-chip and system-in-a-

package solutions are designed to solve
component incompatibilities, reduce
design complexity, and extend product
life and environmental performance in
military applications.

Expand the possibilities;

change your perspective at
www.whiteedc.com/dream.

602.437.1520 TEL 602.437.9120 FAX

www.whiteedc.com

Customer First. Quality Always.

EXPERT ADVICE
Indoor Message System Evaluated
By Börje Forssell
P
ublic and commercial services
such as search-and-rescue,
firefighting, and location-
based services (LBS) require accurate
indoor position determination. Users
face the fundamental problem of
knowing location at the room level,
requiring accuracies on the order of a
few meters. Because of the need for floor
determination, vertical accuracy can be
more important than horizontal.
For indoor applications, stable and
reliable position information means
more than high accuracy.
The United States and Japan have
passed emergenc y-ca ll legislation
requiring third-generation cell phones
to have functions determining their
positions. Rapidly increasing numbers of
cell phones with built-in GPS capabilities
stimulate the LBS market, which is
expected to be the dominating market
for GNSS receivers in the future.
Si g n i f ic a nt proble m s re m a i n,
however. Standard GPS receivers do
not work well indoors, giving position
errors up to several tens of meters — if
a position fix is possible at all. Assisted
GPS (AGPS) improves the situation, but
makes equipment more complicated and
expensive, while reliability and accuracy
may still not suffice. Local solutions
based on RFID, Wi-Fi, ultra-wideband,
pseudolites, and others have been and
6 GPS World | April 2009
are being tried, but with no definitive
solution in sight.
The IMES Concept
The Japan Aerospace Exploration
Agency, GNSS Technologies, and
Lighthouse Technology and Consulting
have proposed and developed an
indoor messaging system (IMES) to
solve the availability issue. Composed
of transmitters, GPS receivers with
modified firmware embedded in cell
phones, and systems of servers, IMES
aims to provide seamless positioning
anywhere in a covered area. Hitachi
Ltd., Systems Development Laboratory
is also working on such a system, as
shown in FIGURE 1.
IMES uses satellite signals outdoors
in the usual way, while using signals
from IMES transmitters indoors, where
satellite signal quality is strongly reduced.
IMES signal structure is similar to that
of GPS satellite signals, except for the
contents of the navigation message. Thus,
the same receiver can be used for both.
An IMES transmitter sends an RF
signal similar to that of GPS and the
Japanese Quasi-Zenith Satellite System
(QZSS), giving its 3-D position, the
Å FIGURE 1 IMES principles, from a Hitachi presentation at Tokyo GNSS Symposium
position of the center of its cell coverage
zone, or linking the receiver to a database
corresponding to an identifier. Instead
of the ephemeris data, clock corrections,
ionospheric parameters, and so on
contained in the GPS message, the
IMES navigation message periodically
dispatches position broadcasts and
additional information in a similar
format.
The concept for seamless transition
between indoor and outdoor use is built
on three further requirements:
䡲 low power consumption,
䡲 low transmitter cost, and
䡲 no harmful interference to other
users of GPS or QZSS.
IMES assumes that positioning
accuracies on the order of 10 meters
will satisfy users, who just want to know
where they are in rooms of moderate
size, shopping areas, parking garages,
and so on. Stable and reliable position
information means more than high
accuracy.
Signals and Codes
IMES uses the same L1 center frequency
as GPS and QZSS, and the same BPSK
modulation. IMES dedicated spread-
www.gpsworld.com
EXPERT ADVICE
spectrum codes are from the same
family of Gold codes as GPS and QZSS
(numbers 173–182 from the C/A-code
assignment table). It is assumed that
signals similar to L1C can be used in
the future. These dedicated codes are
re-used repeatedly but without using the
same code as neighboring transmitters.
The IMES receiver uses the codes only
for de-spreading the spread-spectrum
modulation and as a step to decoding the
navigation message, including transmitter
identification. No pseudorange or time
determination is necessary, because the
desired position is read directly out of the
navigation message. This gives a much
simpler equipment architecture than
when using pseudolites.
Transmitter Power
The transmitted power varies depend-
ing on the location and the distance to
neighboring transmitters. Japanese radio
regulations for license-freesignals require
that the signal level at 3 meters from the
transmitter stay below 35 µV/meter,
limiting the radiated power density from
IMES transmitters.
The radiated power level determines
the coverage zone of a particular IMES
transmitter: larger coverage implies more
power. Because of the requirement for
compatibility with outdoor GPS signals,
IMES power levels received indoors are
expected to be found in the interval
between ⫺126 dBm and ⫺130 dBm
at a distance of about 5 meters from the
transmitter antenna.
The inventors of IMES have conducted
experiments with these power limits in
mind, using transmitter signal levels
from ⫺64 to ⫺76 dBm. The former is
the maximum level allowed for license-
free signals. The working range of IMES
is assumed to be 2–10 meters from the
transmitting antenna, and ⫺64 dBm at
the transmitter means about ⫺120 dBm
at the receiving antenna at a distance
of 10 meters. Accordingly, ⫺76 dBm
transmitted means about ⫺132 dBm at
the receiving antenna.
This latter level is close to what can
be expected for a GPS satellite signal
8 GPS World | April 2009
outdoors and would not create harmful
interference in satellite receivers.
However, the IMES concept foresees
the possibility of having transmitters in
many rooms in large buildings, shopping
areas, and so on, which may imply that
ordinary GPS receivers outdoors in
such an area receive emissions from a
considerable number of transmitters.
The Interference Problem
It is of crucial importance that IMES
signals do not reduce performance for
ordinary GPS receivers used outside the
intended IMES coverage area. Because
IMES uses the same L1 frequency and
the same type of Gold codes as GPS
satellites, the interference will be of
the usual Gold-code cross-correlation
type. If the received power from two
different transmitters is the same, the
cross correlation levels of Gold codes
are between ⫺21.8 and ⫺23.6 dB
(depending on Doppler difference)
below the autocorrelation peak.
The actual level of cross-correlation
interference in a GPS receiver operating
in an IMES environment is determined
by the power received from each IMES
transmitter in question and the number
of these transmitters.
IMES signals are attenuated by walls,
floors, roofs, and other obstacles between
the transmitter and the GPS receiver.
Such attenuation varies between just a
few dB (window or open door) to several
tens of dB (floors/ceilings of reinforced
c oncrete ) . Me a su rement s i n ou r
university building in Trondheim have
given attenuation values such as 20 dB
through a gypsum wall, 35 dB through a
ceiling of reinforced concrete, and 66 dB
through four such ceilings.
Each IMES signal received increases
the cross-correlation interference level in
the GPS receiver. From an interference
point of view, these signals should be
regarded as noise, and the resulting total
variance of these noise-like signals is the
sum of the variances of all the interfering
signals, that is, the sum of the cross-
correlations. Thus, 10 IMES transmitters
in the immediate neighborhood of a GPS
receiver imply a signal-to-interference
ratio up to a maximum of ⫺11 to ⫺14 dB
(referred to the autocorrelation peak),
depending on the transmitted power
and the attenuation situation. Obviously,
in some situations, ordinary reception,
and above all, acquisition of GPS L1
C/A-code signals can be blocked or at
least severely hampered. In addition, a
portable IMES transmitter would make
a good short-range GPS jammer. Using
new GPS signals such as L2C with better
cross-correlation margins would reduce
the problem.
Conclusions
IMES represents a new idea in the field
of position determination in confined
areas. However, the advantages should
not be gained at costs of ordinary
GPS/QZSS users. For this reason,
transmitted power levels should be set
lower than those indicated here and
as proposed by the system developers,
to reduce the probability of harmful
interference. Because of the increasing
availability of high-sensitivity receivers,
it is possible to obtain the wanted
performance with lower power levels.
The IMES group seems to be aware
of this, conducting various experiments
and analyses for optimum setting
and design. We have to wait for final
results, but received power levels around
⫺140 dBm (10 dB below GPS satellite
signals) might still do the job. In all cases,
IMES transmission levels must be
adapted to the local situation. We must
keep in mind, however, that IMES
reception itself is vulnerable to external
interference (including signals from GPS
satellites), so transmitted power cannot
be arbitrarily low.
BÖRJE FORSSELL is professor of navigation in
the Department of Telecommunications at
the Norwegian University of Science and
Technology in Trondheim, Norway.
For reference documents, see the online
version of this article at www.gpsworld.com/imes.
www.gpsworld.com
SYSTEM
THE

Policy and system news and developments | GPS | Galileo | GLONASS

Big Four Race Called from Munich

A debate on the viability of Galileo’s Public Regulated
Service (PRS) simmered just under the surface, and
twice bubbled to the top, at the Munich Satellite
Navigation Summit addressing “The Worldwide Race in
GNSS,” March 3–5. Europe-China tensions were muted
but clearly evident, Russia sounded optimistic notes
about a future conversion to CDMA, and GPS played the
benevolent elder sibling, looking forward to yet more
advance over friendly rivals with the advent of a new L5
signal, in a IIR(M) launch scheduled for March 24.
Galileo Lite. A hint of controversy surfaced in the
opening plenary, when a moderator asked Fotis
Karamitsos, director of the European Commission’s
(EC’s) Maritime Transport, Galileo & Intelligent Transport
unit of the Directorate-General for Energy and Transport,
about the prospects for a “Galileo Lite” system of
20 satellites broadcasting only an Open Service signal.
Karamitsos emphatically refuted the idea, stating that
“since July 2008 a solid budgetary and legal base for
the fully defined Galileo system and EGNOS has been
established with strong support from the European
Parliament and member states. Member states have a
legal obligation to deliver the full system. They all signed
the regulation, the strongest legal instrument that we
have in the Community.
”If all goes well, we will be able to conclude the
contracts for most of the [seven] work packages
starting from the middle of the year,” Karamitsos added.
UK Transport. The issue lay relatively quiet until
the last morning, in a session aptly titled “GNSS Hot
Topics.” Anne Sta, Head Galileo Programme Division
Department for Transport, United Kingdom, stated that
in the UK, “It is difficult for us to see a need for the
PRS. I struggle with the military’s need for this system,
because it would never be under their control. It would
be under control of the civil sector.”
She cited two studies commissioned from
consultants, at least one of which, she promised, would
be posted by the end of April on the transport agency’s
website, regarding cost considerations and likelihood
of user adoption of the PRS. She claimed that the UK
military has no need of it, and that consultants were able
to identify fewer than five other users — not groups, she
specified, but actual users — of the PRS.
Richard Peckham, business development director for
EADS Astrium in the UK, added in a later presentation
that “PRS is very strongly French today.” To the UK
industry, he suggested, “It’s not clear that PRS is
necessary, practical, or cost-effective for emergency
services. From my point of view, it’s probably not.
Speaking for industry, we’re looking for clarity. Who
really will be the users?”
Overlay. In other conference proceedings, it appeared
that the EC and China have not resolved an impasse
regarding overlay of proposed Compass signals on two
L1 frequency bands planned for Galileo’s same PRS.
“There needs to be spectral separation so there is no
interference with government services,” reiterated Paul
Verhoef, head of unit, Galileo and intelligent transport,
for the EC. “China needs to abide by ITU (International
Telecommunications Union) allocations.”
“There are a couple of new elements in our
relationship to be worked on,” Verhoef added.
“A Technical Working Group on compatibility and
interoperability, and an Industrial Working Group. We are
looking for recommendations by the middle of the year.”

DAGR Extended
From a Space & Missile Systems
Center media release, March 17:
“The GPS Wing awarded a
sole-source follow-on production
contract for Defense Advanced
GPS Receiver with a ceiling of
$450 million to Rockwell Collins Inc.
of Cedar Rapids, Iowa, March 10.
“The first production option
buy of $87 million will be put on
10 GPS World | April 2009
contract by the end of the month.
This will provide for an additional
51,200 selective availability anti-
spoofing module-based DAGRs
and associated accessories. This
represents a $200 savings per unit
over the previous contract.
“‘I am extremely pleased with the
outstanding accomplishments of
both parties to negotiate a win-win
agreement,’ said Col. Dave Madden,
commander, GPS Wing. ‘The
warfighter will directly benefit from
their success, and DAGR will continue
to provide real-time PNT services to
the fingertips of the soldiers fighting
the ongoing war on terror.’
“Under the current contract,
Rockwell Collins has delivered more
than 290,000 DAGRs to the armed
forces of the United States and
22 Foreign Military Sales customers.
DAGR is the current state-of-the-art
military handheld GPS receiver ...
The DoD has recognized the
(Continued on page 12)
www.gpsworld.com
THE SYSTEM
Jamming Trial Confirms Vulnerabilities
Coast Guard, Executive Branch Kill eLoran Nonetheless
The General Lighthouse Authorities
of the United Kingdom and Ireland
conducted a GPS jamming exercise
in 2008 to investigate the effect
of GPS service denial on maritime
navigation and DGPS monitoring
operations, as well as evaluate the
performance of eLoran.
The trial concluded that GPS ser-
vice denial has a significant impact
on maritime safety. eLoran was un-
affected by GPS jamming and dem-
onstrated an accuracy of 8.1 meters
(95 percent), which is comparable to
stand-alone, single-frequency GPS.
“In the future e-Navigation envi-
ronment, the combination of GPS,
Galileo, and eLoran will provide
robust and resilient PNT in order to
reduce the impact of human error
and to improve the safety, security,
and protection of the marine environ-
ment,” the report concluded.
The exercise and its results are
described in GPS Jamming Trial
Summary Report (Report No. RPT-
26-AJG-08). Find the Executive
Summary at www.gpsworld.com/maritime.
Washington Acts. In February, the
Office of Management and Budget
announced the President’s Fiscal
Year 2010 Budget. In the section for
the Department of Homeland Se-
curity (DHS), the budget “supports
the termination of outdated systems
such as the terrestrial-based, long-
range radionavigation (LORAN-C)
operated by the U.S. Coast Guard
12 GPS World | April 2009
(USCG), resulting in an offset of
$36 million in 2010 and $190 million
over five years.”
In a message to USCG personnel,
Vice-Admiral V. S. Crea reiterated
that the budget sent to Congress
calls for termination of Loran-C, cit-
ing the quoted passage. Since 1994,
Loran has been regularly rescued by
Congressional advocates, but the
current economic climate may make
such restored funding.
Has Loran reached the end of its
road?
Government-commissioned tim-
ing criticality studies, a white paper
on GPS backup prepared under the
auspices of the Federal Aviation
Administration, and an Independent
Assessment Team headed by Brad
Parkinson, program director for the
GPS Joint Program Office at GPS
inception, have all endorsed eLoran
as a needed back-up for GPS in the
event of attack or other disruption.
( DAGR, Continued from page 10)
importance of DAGR and has
provided millions of dollars of
supplemental funding to get
these GPS receivers into the
hands of the warfighter as quickly
as possible.” [end official release]
Unofficial Word. Characterizing
the currently configured device
as obsolete, one source stated
“This equipment will sit on the
shelves and never be used.”
Military personnel in the
combat theater widely use
commercial GPS receivers and
PNDs in addition to, or some-
times instead of, government-
issue GPS equipment.
The figures in the second
paragraph of the SMSC release
calculate out to a cost of
$1,699.21 per handheld receiver.
Briefly Noted
GLONASS authorities changed the
operating frequency channels of
two pairs of antipodal satellites on
March 11, with no apparent advance
notice or explanation. GLONASS 717
in orbit slot 10 was changed from
channel 4 to channel -7; 715 in orbit
slot 14 was changed from channel 4
to channel -7; 718 in orbit slot 17 was
changed from channel -1 to chan-
nel 4; GLONASS 725 in orbit slot 21
was changed from channel -1 to chan-
nel 4. Modern GLONASS and GPS/
GLONASS receivers should have
handled the transition seamlessly.
The Lockheed Martin team develop-
ing GPS III (Lockheed Martin Space
Systems, ITT, General Dynamics)
successfully completed 19 of 71
PDRs for key GPS III spacecraft
subsystems and assemblies as part
of the Preliminary Design Review
(PDR) phase with the U.S. Air Force.
Completed steps include L-Band
transmitters, antennas, solar arrays,
power regulation unit, all attitude
control assemblies, and tracking
telemetry and command subsystem
and assemblies. The effort will
culminate in an overall GPS III
Segment PDR in May to ensure
the preliminary design meets
requirements prior to advancing to
the Critical Design Review phase.
The GPS Wing has scheduled its
2009 GPS Partnership Council
meeting at the Space and Missle
Systems Center at Los Angeles Air
Force Base for May 5–6.
The Wing, federal agencies, and
industry partners will discuss
“Leading Space-Based PNT for the
Warfighter, the Nation, the World.”
Registration is open at www.
afcea-la.org/events/gps2009/. A scramble
golf tournament (not included in con-
ference fee) takes place Wednesday
during the Warfighter panel.
www.gpsworld.com
 BUSINESS
THE
Industry news and developments | GPS | Galileo | GLONASS
» MILITARY & GOVERNMENT
Security Paramount, Strengthened by Location
Two new products, from Qascom
in Italy and White Electronics in the
United States, demonstrate ways to
protect and secure sensitive data,
augmenting other encryption and
protection with location technology
in scalable security solutions.
Qascom’s TIGER project involves
a security platform using GNSS. The
concept acts through an access con-
trol token, within the cost constraints
of mass-market applications. The
architecture spans signal quality
monitoring to internal processing of
the receiver to develop signal
authenticity and quality validation
methods, such that it can determine
the degree of trust of a navigation
solution. Through TIGER,
䡲 employees can access sensitive data
only inside a specified geographical
area;
䡲 an e-mail can be decrypted only in
predetermined locations;
䡲 a stolen laptop stolen would not
reveal any secret information if taken
outside a predefined zone;
䡲 critical operations can take place only
inside a predetermined zone;
䡲 proof of position and contextual
data, such as time, can be included
in the digital signature of an e-mail,
proving where and under what con-
ditions the e-mail was written;
䡲 managers can analyze in real time
the locations from which employees
or customers are accessing the enter-
prise network.
The TIGER concept targets both
new and existing security applica-
tions with features that range from
a security scheme that grants per-
mission to access and manipulate
information based on the physical
location; geoencryption based on lo-
cation and time, so that data can be
14 GPS World | April 2009
p QASCOM’S Trusted Innovative GNSS rEceiveR (TIGER) concept
accessed (decrypted) only in autho-
rized areas; and proof of the position,
time, and location acquisition context
of a given person or entity.
Qascom plans to offer TIGER in a
USB device connected to a PC, USB
for virtual private network (VPN)
access control, memory card for
storage of geoencrypted data, and
personal devices such as mobile
phones or PDAs.
Military Security. In language
clearly addressing military scenarios,
White Electronics states that “Be-
cause data security relies on the
execution of a key or keys, which
if found out, would void all protec-
tion from effectiveness, it places a
human dependency in the middle of
the security matrix. Those protecting
valuable data might refuse to give up
that critical information even under
the most heinous duress, but the
potential for that breach exists.”
The company’s ZoneLoc acts
across multiple formats including
compact flash, USB, solid state
device (SSD), or other. Using an RF
signal to establish a proximity bound-
ary, ZoneLoc defines an area or zone
within which data will be contained
by a specified data storage device.
If this zone is broken, the device au-
tomatically purges the data without
human intervention.
A protection zone can include
proximity and a pre-established set
of parameters, or features tailored
to a particular application. An ap-
plication may define the zone by
physical proximity and an associated
hardware device. Another applica-
tion may adapt to a mobile data
storage device, restricted by a larger,
definable physical boundary, from a
conventional secure military area to a
larger theater of operations.
White offers a demonstration kit
including a CF card and a wireless
key device that defines the center
of the zone. ZoneLoc complies with
specifications of: NISPOM DoD
5220.22-M, NSA-130-2, Air Force
AFSSI-5020, Army AR380-19, and
Navy NAVSO P-5239-26, all stan-
dards that govern sanitization
requirements by most agencies
in the military market.
White claims that “this may be
the smallest form factor of SSD ever
developed that complies with these
erasure standards.”
www.gpsworld.com
THE BUSINESS
» MASS MARKET OEM
Broadcom, SiRF Top ABI’s GPS Chip Vendor Ranking
Broadcom has been ranked at the
top of the latest Vendor Matrix
released by ABI Research. SiRF
and Texas Instruments claimed
the second and third spots in the
company’s most recent evaluation
of worldwide GPS chip vendors.
”Recent consolidation within
the GPS chipset sector has
led to intensified competition
among the remaining players and
increased collaboration between
semiconductor suppliers and the
firmware and complementary radio
technology providers (primarily
Wi-Fi and Bluetooth),” said ABI
Research senior analyst George
Perros. “Sector participants are
attempting to differentiate their
product offerings and re-tool their
value propositions, particularly
in consumer device markets. The
result is a GPS semiconductor
sector that is evolving into a
two-tier industry, where smaller
participants will find it difficult if not
» SURVEY & UTILITIES
Getac GPS Comes with Camera, Altimeter, E-Compass
Rugged GPS maker MiTAC Technol-
ogy Corporation of Telford, England,
is introducing the Getac PS535F, a
GPS personal digital assistant with
an autofocus camera, altimeter,
and E-compass. Its features are de-
signed so field-based
employees can cap-
ture field data on
site and transmit
data back to head-
quarters in real-
time, as well as
provide pinpoint
navigation sup-
port to improve
location-based
productivity.
The PS535F
is designed
16 GPS World | April 2009
GPS Integrated Circuit
Manufacturer Vendor
Matrix Top 10 Listing
1. Broadcom Corporation
2. SiRF Technology Inc.
3. Texas Instruments Inc.
4. Qualcomm Inc.
5. St Microelectronics
6. u-blox AG
7. Atheros Communications Inc.
8. Infineon Technologies AG
9. Atmel Corporation
10. MediaTek Inc.
impossible to compete for large
volume applications.”
The Vendor Matrix is an analytical
tool developed by ABI Research
to provide a clear understanding
of vendors’ positions in specific
markets. Vendors are assessed
on the important parameters of
innovation and implementation
across several criteria unique to
each vendor matrix, the research
company said.
for GIS surveying applications,
including utilities, oil and gas, forest
patrol, and surveying and mapping
of geology and mineral resources.
The built-in 3 megapixel auto-focus
camera enables instant visual data
capture in the field, the company
said, and the VGA touchscreen
display with sunlight readable
technology enhances the viewing
quality of graphic-intensive files and
detailed maps.
To enhance its navigation
support, the PS535F has a built-
in altimeter and an E-compass,
allowing pinpoint positioning of
longitude, latitude, and altitude, as
well as providing the directional
relation between the targeted
subject and the user’s position.
For this matrix, under innovation,
ABI Research examined design
innovation, the use of evolving
fabrication technology, packaging
innovation, features (software/
firmware, power, size), flexibility
(chip packaging, OEM integration
options), the vendors’ recent
product introductions, and whether
the integrated circuits offer multiple
radio integration.
Under implementation, ABI
Research looked at the ability of
the design to meet performance
and feature objectives; the
vendors’ sourcing strategies and
the strength of their supply chains;
marketing/sales strategy, structure,
and successes; organizational
alignments and structures to
support GPS market strategy;
market presence; the effectiveness
of product strategies and strength
of product lines; partnerships;
relative market share, and overall
organizational health.
» MILITARY
Raytheon to Produce
Anti-Jamming Antenna
Raytheon Systems Limited (RSL)
has received a U.S. Air Force
contract for initial production of the
GPS Advanced Digital Antenna
Production system. The ADAP
system protects GPS-based navi-
gation and precise timing systems
from deliberate jamming and
accidental interference.
The $1.4 million contract from the
GPS Wing provides for 41 antenna
electronics and 28 controlled recep-
tion pattern antenna to equip U.S.
Navy surface platforms and fixed-
and rotary-wing aircraft. It is a suc-
cessor to the GAS-1 GPS antenna
system that has been in continuous
production at RSL since 1998.
www.gpsworld.com
 THE BUSINESS

» SYSTEM DESIGN & TEST » MILITARY & GOVERNMENT

ITT Awards Symmetricom Timing Contract for GPS III Rugged Notebook
Symmetricom, Inc. has been
awarded a contract from ITT Corpora-
tion Space Systems Division valued at
a minimum of $4 million, part of the
building of GPS IIIA, the next-genera-
tion GPS Space System program.
Under the award, Symmetricom will
deliver its Model 9552 Ovenized
Quartz Master Oscillators over three
years, including developing, qualify-
ing, and delivering flight oscillators for
the first two space vehicles.
“The Model 9552 oscillator con-
tinues Symmetricom’s legacy in the
GPS program beginning with our
cesium clocks for GPS blocks I, II,
IIA and IIF,” said Peter Cash, director,
Space, Defense & Avionics Group at
Symmetricom.
The Model 9552 is a high perfor-
mance oven-controlled quartz crystal
oscillator that produces a highly
stable, low noise reference frequency
output, Symmetricom said. It is
suited to space applications and can
be used as the on-board frequency
reference for a satellite or ground sta-
tion. The oscillator will produce the
output frequency for the GPS IIIA
Navigation payload by means of
steering from the on-board Atomic
Frequency Standards.
The U.S. Air Force announced in
May 2007 that Lockheed Martin was
awarded the contract to build GPS III.
Under the $1.4 billion contract,
Lockheed Martin along with ITT
and General Dynamics will produce
the first two GPS IIIA satellites with
first launch projected for 2014. The
contract also includes options for up
to 10 additional spacecraft.
The new GD8000 notebook
computer by General Dynamics
Itronix is now available for military
personnel, field service workers,
and first responders who require
high-performance computing while
working in extreme temperatures,
humidity, dust, and rain-soaked
environments. GD8000 features
include integrated GPS and wireless
connectivity with Wi-Fi (WLAN),
optional high-powered WLAN and
Bluetooth; weighs 7.9 pounds
fully equipped.

THE CHALLENGES OF MAPPING URBAN HOLLAND.

AS SEEN BY NOVATEL’S GPS+INS TECHNOLOGY.

Moerdijk, Oostflakkee or wherever; you can now map cities as if you’re right there. Thanks to
Cyclomedia’s 360-degree panoramic digital mapping with NovAtel precise positioning technology
onboard, the world appears in 3-D because, well, it is. Incorporating NovAtel’s SPAN GPS+Inertial
Measurement Unit system, accurate and robust positioning data is possible even in challenging urban
conditions where GPS alone is limited.

3-D mapping is just one of many worldwide applications successfully utilizing NovAtel’s precise
positioning products.

To learn more about the products that powered this application,

navigate your way to novatel.com.

NovAtel’s precise thinking makes it possible.

1-800-NOVATEL (U.S. & Canada) or 403-295-4900 | Europe +44 (0) 1993 852-436 | SE Asia & Australia +61 (2) 8668 4073 | sales@novatel.com

www.gpsworld.com April 2009 | GPS World 17

THE BUSINESS

» TRANSPORTATION

Navteq Reports Continued Growth for In-Vehicle Nav

2008 was a year of continued
growth for the in-vehicle navigation
industry, with the introduction
of navigation as a new option on
26 North American car models
across 15 automotive brands,
according to map provider Navteq.
Brands included Aston Martin,
BMW, Chevrolet, Dodge, Ford,
Honda, Hummer, Hyundai, Infiniti,
Jaguar, Kia, Lincoln, Nissan, Toyota
and Volkswagen. In 2008, 208 car
models offered in-vehicle navigation
out of the 318 models sold.
As the percentage of vehicle
models offering navigation climbed
to 72 percent in 2008, so did general
awareness of navigation, according
to the company. Navteq’s annual
consumer awareness tracking study
shows that in 2008, awareness of
navigation approached 100 percent,
with 94 percent of the U.S. driving
population considering themselves
familiar with navigation, versus
84 percent in 2005.
In addition, real-time traffic infor-
mation for in-vehicle navigation more
than doubled from 2007 to 2008,
with a total of 94 car models offer-
ing traffic information. Overall, the
installed base for vehicles sold with
navigation reached 8 million in 2008.
“In 2008, 56 percent of the U.S.
driving population stated that in-
vehicle navigation is somewhat or
very important in their next vehicle
purchase decision,” said Dean Von-
Bank, Navteq’s vice president and
general manager — Vehicle Ameri-
cas. “Even in a struggling economy
we expect increased consumer inter-
est in navigation and are optimistic
consumers will continue to demand
in-vehicle navigation in 2009.”

» PROFESSIONAL OEM

Value of GPS Networks, Services Predicted to Rise

The worldwide value of GPS net- and demand for precision GNSS “The adoption of precision GNSS
works and services is forecast to rise Augmentation Infrastructure and products and services, particularly
from US $263 million in 2009 to US Services, including CORS, RTK for machine control, are reaping
$504 million by 2013, according to a Networks, RTK Clusters, SBAS, impressive productivity gains for
new report by Position One Consult- and DGPS. users”, said Lorimer, managing
ing. The report is authored by Eric At the core of the 130-page director of Position One Consulting.
Gakstatter, who is also GPS World’s report are five-year growth and “Many are opting to consume pre-
survey and construction editor, and financial projections with optimistic, cise positioning services rather than
Rob Lorimer, GPS World’s profes- realistic, and conservative scenarios operate their own reference stations,
sional OEM editor. to reflect current macro economic and this report analyzes that trend.”
The report includes global trends uncertainties. The report covers background
and status for the entire spectrum of
» EVENTS GNSS, including GPS, GPS modern-
ization, GLONASS and GLONASS
GPS-Wireless 2009 modernization, SBAS (Satellite-
www.gps-wireless.com Based Augmentation Systems), Gali-
San Francisco, California: April 28–29, 2009 leo and Compass/Beidou, QZSS and
This annual event brings executives together to discuss important and IRNSS, as well as technology/market
converging navigation, location-based services, fleet and asset manage- trends and potential disruptions.
ment for enterprise markets, online portal, personal location, and home- The report was introduced at
land security markets. the Munich Satellite Conference,
March 3-5, 2009. E-mail
European Navigation Conference — GNSS 2009 abstracts@positiononeconsulting.com for
www.enc-gnss09.it an abstract.
Naples, Italy: May 3–6, 2009
Exchange news and information about progress in the radio-navigation
MORE ONLINE
sector, with emphasis on satellite, ground segment, and user segment
technologies in the frame of the Global Navigation Satellite System. In-depth Coverage
Find more details on these stories
at gpsworld.com. For the latest GPS
See all events at www.gpsworld.com/events. news, sign up for GPS World Alerts.

18 GPS World | April 2009 www.gpsworld.com

BLACK YELLOW MAGENTA CYAN GPS0409_019-020.pgs 03.19.2009 13:53 janinej
BLACK YELLOW MAGENTA CYAN GPS0409_021-022.pgs 03.19.2009 13:52 janinej
BLACK YELLOW MAGENTA CYAN GPS0409_019-020.pgs 03.19.2009 13:53 janinej
BLACK YELLOW MAGENTA CYAN GPS0409_021-022.pgs 03.19.2009 13:52 janinej
 𰂇 2 different mounting
options:
female thread 1”-14
3 holes M5 50

BLACK YELLOW MAGENTA CYAN GPS0409_025-026rev.pgs 03.19.2009 14:14 janinej

BLACK YELLOW MAGENTA CYAN GPS0409_023-024.pgs 03.19.2009 13:54 janinej
BLACK YELLOW MAGENTA CYAN GPS0409_023-024.pgs 03.19.2009 13:54 janinej
 𰂇 2 different mounting
options:
female thread 1”-14
3 holes M5 50

BLACK YELLOW MAGENTA CYAN GPS0409_025-026rev.pgs 03.19.2009 14:14 janinej

BUSINESSOUTLOOK
How the Stimulus Plan Benefits Our Industry
T
he plan is in place. How will the American Recovery
and Reinvestment Act of 2009 affect our industry?
Is it good news or bad?
The answer is, you have to dig for bright spots, but they’re
there. The economic plan that has been signed into law has a vast
majority of backers as well as quite a few pundits, especially on
Wall Street. Regardless of the current situation — the spending
authorized on the recovery act will begin taking place through
a miracle of various programs and infrastructure projects —
consumer spending is at an all-time low.
The GNSS community is widely varied as it relates to
different applications and product types. To overlay the
$787 billion plan with the industry, one must understand
which programs will spur growth and sales based on category
of GPS use. Unfortunately, the consumer PND mass market
side will take the longest to reap any rewards, but other sectors
will gain from the windfall initiatives outlined in the plan. The
original request for authorization was targeted at $819 billion
but negotiated down to its current level.
Recently a chart began showing up on the Internet based on
the Department of Commerce projections. Initially the bud-
get was based on the $819-billion plan. Although the dollar
amount is less ($787 billion) it’s an easy way to visualize the cat-
egories and trickle-down effect of the various programs which
will stimulate the GPS industry. Looking at the categories of
From The Washington Post
www.gpsworld.com
investments, I can see more than a
few areas that will shore up sagging
balance sheets for GPS manufactur-
ers in the times ahead:
Rural Development. High on the
list is $11.1 billion for development
of rural broadband networks. This
will spur GPS survey and GIS data- By Stephen Colwell
collection sales in new equipment Editor, Mass Market
purchases, upgrades, and training. OEM Newsletter
Another $4.1 billion is earmarked
for rural development, which probably means infrastructure
projects related to construction and network development.
Commerce and Justice. $2.8 billion will be deployed to
expand existing broadband networks, which will in turn result
in increased activities for land surveyors, engineers, right-of-
way consultants, and equipment manufacturers. $3 billion has
been set aside for law enforcement agencies to increase hiring
and equipment. PND sales could increase their market share
in this area.
Energy. $18.5 billion has been set aside for renewable energy
projects. This will lead to the increase of wind and solar energy
farms. All phases of mapping, GPS, and GIS, along with
surveying and civil engineering, will profit from this sector.
$8 billion will go to loan guarantees, renewable energy systems
(RES), and electricity transmission.
Here we begin to see the start of a $17.4-billion investment on
development of one of the largest infrastructure projects, which
is upgrading the nation’s antiquated power grid. Army Corp will
receive an additional $4.5 billion, which will probably be used
to continue work building levies and flood-control systems.
This sector as it unfolds through various programs will be a huge
boost to the GPS and GIS manufacturers. Mapping-related
services will also see a bright spot.
There are quite a few pots that manufacturers of mapping
and GPS systems can target. It’s an opportunity where the risk
factor for trying new technology is offset by federal funding.
This means that existing manufacturers and providers of
mapping-related projects will see a definite increase in their
sales numbers; it also opens the door for new competition.
Price, terms, product worthiness, and company support
programs will all go a long way towards increasing the dollars
that flow into the GPS industry.
There’s more! Read the full column from the March Mass Market
OEM Newsletter at www.gpsworld.com/stimulus.
April 2009 | GPS World 27
TRANSPORTATION | Vehicle Integration
Positioning for Driver Assistance
Communication on the Road
A low-cost GPS-based positioning
architecture for V2V and V2I systems
uses a scalable, flexible system
with over-the-air messaging
and an onboard design
to enable a range of
in-vehicle safety
features.
Chaminda
Basnayake,
General Motors
Research and
Development
p FIGURE 1 Sensing with traditional sensors (left) and V2V 360-degree awareness (right)
V
ehicle-to-vehicle (V2V) communication-based ad-
vanced driver assistance systems (ADAS) enable or
enhance a suite of safety features by combining GPS
and short-range wireless communication. The allocation of
the 5.9 MHz radio spectrum for highway safety and traffic
efficiency applications using Dedicated Short Range Com-
munications (DSRC) has given automotive manufacturers the
ability to go beyond traditional sensor-based safety features to
provide drivers with another sense. DSRC enables vehicles to
talk to surrounding vehicles and maintain a comprehensive
situational awareness of roadway traffic.
General Motors first publicly demonstrated a fleet of com-
municating vehicles with driver warnings and automatic vehicle
braking in 2005. Recently, General Motors has further developed
the system with its OnStar and StabiliTrak technologies. The
latest GM V2V fleet is fully integrated with production tech-
nologies with the exception of DSRC. Its built-in communica-
tion system can also interface with roadway infrastructure and
support a range of new vehicle-to-infrastructure (V2I) features
as they become available.
An accurate, reliable, and affordable positioning system, flex-
ible enough to evolve with emerging V2V and V2I application
requirements, obviously constitutes a prerequisite for these sys-
tems to work. This system is based on an efficient over-the-air
(OTA) messaging scheme and an onboard system design that can
be scaled to fit the needs of a range of in-vehicle safety features.
As demonstrated in field tests, the system could support Which-
Road, Which-Lane, or even Where-in-Lane applications using
the same basic architecture.
Why V2V? Location and situation awareness are key enablers
for almost all modern-day vehicle safety and convenience
systems. These include GM OnStar, adaptive cruise-control
28 GPS World | April 2009
(ACC) systems, side blind-zone alert systems, and many more.
Most of these applications require specific knowledge about
vehicles in close proximity, typically their speed, acceleration,
and heading.
In an ACC implementation, radar, lidar, or other technologies
track target vehicles in the field of view. However, this requires
multiple sensors to cover different regions around the vehicle and
to address the range requirements of different applications. For
instance, forward- or backward-looking safety applications typi-
cally require the pairing of short- and long-range radar sensors
and separate sensors to provide coverage of side blind zones as il-
lustrated in FIGURE 1 (left panel), showing tracked and untracked
vehicles using dark and light shades respectively.
These traditional sensors do not provide coverage beyond im-
mediate line-of-sight, which may limit the effectiveness of these
applications. For instance, in a hard-brake event that occurs far
ahead, following drivers may not see the event until the vehicle
directly ahead takes a responsive action. V2V systems can provide
instantaneous communication between vehicles, even in situa-
tions where an event is obscured by a sharp turn, a large vehicle,
or some other obstruction.
If vehicles carry location awareness technology such as GPS
and the ability to talk to each other, all vehicles and even the
infrastructure elements such as traffic-control devices could
communicate and incorporate situation awareness onboard. In
Figure 1’s right panel, each vehicle can build an awareness zone
(AZ) using data received from everyone within its communica-
tion range, and provide feedback to the driver. This concept is
central to V2V communication.
V2V technologies eliminate two critical limitations of tra-
ditional sensor-based technologies. First, a single communica-
tion sensor can communicate with enabled vehicles within its
www.gpsworld.com

communication range, making 360-
degree situational awareness possible with
only a single sensor. Second, communica-
tions range is not limited by line-of-sight,
but only by the transceiver’s range. A
single sensor enables the formation of an
AZ around the vehicle, and all equipped
vehicles in the AZ can be tracked (shown
in darker shade in Figure 1’s right panel).
Further, communications between ve-
hicles serve as a medium for exchange of
a host of other information. For example,
traditional sensors rely on analog sensing
to indicate the presence of a vehicle and to
estimate certain limited target data such
as instantaneous speed, acceleration, and
heading. V2V systems rely on digital com-
munication packets that could send this
and other specific data relating to future
actions, such as turn-signal status and
subsystem data such as stability status.
Why V2I? This medium also enables
information exchange with communica-
tion-enabled infrastructure entities such
as traffic-control devices or traffic signs.
Thus V2V can extend to include V2I
using the same technologies. In essence,
once a vehicle is within the communica-
tion range of an equipped intersection, it
becomes aware of all the information sent
by all the vehicles around itself and the
infrastructure elements. V2X collectively
identifies the common underlying system
components of V2V and V2I.
Of the two critical enabling technolo-
gies — GPS and DSRC — GPS is closer
to becoming a standard technology in
most vehicles today. GM OnStar, which
includes GPS capability, is already avail-
able in more than 50 GM models. The
Federal Communications Commission’s
allocation of the 5.9 MHz band for high-
way safety and efficiency applications
using DSRC is a commitment to estab-
lishing the second key enabler. Tests show
that vehicle communication using DSRC
is reliable up to a range of around 300 me-
ters in typical roadways, sufficient to meet
the requirements of ADAS applications.
ADAS Applications
The GM V2X demo fleet carries a range
of state-of-the-art displays and actua-
www.gpsworld.com
Vehicle Integration | TRANSPORTATION
p FIGURE 2 V2V ADAS dashboard interface (left); illuminated icon in side mirror (right)
p FIGURE 3 ADAS icons: top row, left to right, SVA, EEBL, FCW, and LCA. The FCW icon
shows all three possible states of the feature of which only one state is shown at a time.
The green indicates a vehicle detected in front, and the amber and red correspond to two
warning levels of a potential rear-end crash. The bottom row depicts cross-path vehicle
ahead, slippery road ahead, and pedestrian ahead warnings.
tors, including high-head-down displays LCA monitors the state of vehicles
(HHDD), mirror icons (see FIGURE 2), in adjacent lanes and displays informa-
haptic seats, and automatic braking. tion — an illuminated icon in the side
The f leet can demonstrate slow/ mirror, see Figure 2, right — when it de-
stopped vehicle ahead advisor (SVA), tects a communicating vehicle in the host
emergency electronic brake light (EEBL), vehicle’s blind spot. It also alerts the driver
forward collision warning (FCW), and when a lane-change maneuver is unsafe
lane change advisor (LCA), among many due to an equipped passing vehicle.
other ADAS applications. FIGURE 3 depicts several HHDD and
EEBL identifies situations in which a side-mirror icons. The cross-path vehicle
vehicle ahead of the host is braking hard. ahead shown in the bottom row could be
It extends the driver’s view by providing supported simply based on V2V capa-
important information that may not bilities. The slippery road ahead applica-
be visible to the driver because of other tion requires input from vehicle stability
vehicles between the braking vehicle and systems such as StabiliTrak. Equipping
the host. Based on braking intensity and a pedestrian with a portable device that
other information such as lane orientation, communicates with onboard V2X sys-
EEBL triggers an audio tone and visual tems extends the system to pedestrians.
warnings to attract driver attention.
Since the information is communicated Positioning Technologies
instantaneously, drivers have more time V2X ADAS applications vary widely in
to respond appropriately. their position accuracy requirements.
In FCW, relative distance, relative ve- FCW requires higher accuracy position-
locity, and relative acceleration identify ing information since it must identify
potential rear-end crash scenarios. Upon same-lane vehicles. EEBL only requires
detecting such an event, the system actu- identification of vehicles in the same
ates visual and audio displays and the front road. Some applications need accurate
portion of a haptic seat to attract driver at- positioning information in an absolute
tention. The fleet can also automatically sense, whereas others may function with
brake in situations where the driver does positioning information accurate in a
not react to an imminent collision. relative sense. For instance, almost all
April 2009 | GPS World 29
TRANSPORTATION | Vehicle Integration
p FIGURE 4 Scalable onboard processing architecture
V2V features may be enabled by know-
ing the relative location and orientation
of the surrounding vehicles (that is,
across- and along-distance to all target
vehicles), although knowing the abso-
lute information about the vehicles with
respect to a map is preferable.
Even a feature such as intersection
cross-path warning can be enabled with
relative location and orientation infor-
mation. However, almost all V2I features
require absolute positioning information
so that roadway features are properly ori-
ented with respect to vehicle location.
Typical example is the stop-sign warning
feature that could warn the driver of po-
tential stop-sign violations.
GM R&D uses this positioning accu-
racy classification:
䡲 Which Road: > 5 meters
䡲 Which Lane: > 1.5 meters
䡲 Where in Lane: > 1 meter
This classification is used in expressing
both absolute and relative accuracies. A
given feature may belong to two accuracy
ranges in absolute and relative accuracy
requirements. For exa mple, LC A
requires Which-Lane or better relative
accuracy, while the absolute accuracy
could be Which-Road or better. GM
has done extensive research on different
GPS-based positioning technologies,
including real-time kinematic (RTK),
that enable Which-Lane and Where-in-
Lane-level positioning.
Real-World Challenges
Real-world success of ADAS depends
30 GPS World | April 2009
on a minimum in-vehicle technology
set, including communication medium,
GPS, and a standardized messaging
format. Industry-wide standards and
technology penetration are therefore
prerequisites for successful V2V sys-
tems, while infrastructure support is an
added prerequisite for V2I. All vehicles
must communicate using the same
communication technology and OTA
messaging structure for interoperability.
The OTA messaging structure used by
the GM demo fleet closely resembles the
SAE J2735 standard message set under
development. The positioning system
architecture presented here answers two
key real-world challenges:
䡲 Efficient and flexible OTA
positioning data exchange
䡲 Scalable onboard resource use.
Data Exchange. Data exchanged
between V2X entities can range from
current and historical vehicle status to in-
tersection maps and control status infor-
mation. Given that there could be several
hundred vehicles within communication
range of any vehicle and the necessity for
frequent communication for time-critical
applications, efficient use of the commu-
nication bandwidth assumes paramount
importance. While communication strat-
egies such as power and rate control may
address part of this, the positioning sys-
tem design process itself can contribute
significantly. The architecture given here
is designed with this requirement in mind,
using a two-stage OTA positioning-data-
sharing approach, such that positioning-
related data volume is minimized.
The heartbeat message contains:
䡲 a GPS time tag
䡲 vehicle position (lat, long, elevation)
䡲 vehicle kinematic data (speed, head-
ing, acceleration, yaw rate)
䡲 a set of other vehicle state informa-
tion essential for the applications.
As the name implies, this part of the
OTA message is transmitted on a fixed-
time interval as required by the safety ap-
plications. Position information is used to
estimate Which-Road-level absolute and
relative position of the host and other V2X
entities, just by decoding the OTA data,
without the need for further processing
using DGPS or RTK.
The second part of the message con-
tains predefined optional data that can
be attached to a heartbeat message; one
message type contains data necessary for
RTK positioning. The heartbeat message
contains a list of attached optional data,
its type identified by a MessageType in
the optional dataset. This is followed by
a precise-positioning data header and the
data packaged in a predefined format.
The current GM fleet packages RTK data
with the RTCMv3.0 format. Though not
customized for automotive applications,
it was considered the most reliable and
efficient existing format for this purpose.
These data enable any given vehicle to
establish an RTK relative position vec-
tor between itself and any other entity
sending RTK-capable GPS raw data,
thus enabling Which-Lane or Where-
in-Lane positioning, depending on the
hardware support. This approach allows
vehicles to send GPS raw data attached
only to selected messages. For instance, a
vehicle could send out heartbeat messages
at a rate of 10 Hz and the raw RTK data
at 2 Hz by attaching the optional GPS
raw data to every fifth heartbeat message.
While reducing the transmission rate of
GPS raw data could affect an RTK solu-
tion such as convergence time, this can
become a significant saving in real-world
operation when hundreds of vehicles are
communicating with each other. In es-
sence, the stage-based approach gives
the flexibility to reach a balance between
www.gpsworld.com

precise positioning performances against the communication
bandwidth usage.
Scalability. Although the processing capabilities and the num-
ber of processors onboard vehicles have dramatically increased
over the last decade, user applications and their computational
needs have grown at an even greater rate. Because of associated
cost, processing requirements of the relative positioning system
must be carefully managed so that it becomes affordable.
This relative positioning architecture addresses the optimal re-
source management need by enabling it to function with minimal
processing dedicated to Which-Road-level positioning. At the
same time, it provides the flexibility to enable Which-Lane-level
accuracy relative positioning for all or for a subset of communi-
cating entities, depending on processing resources available.
A vehicle with minimal onboard processing may use only the
heartbeat message data and enable Which-Road applications.
Another could use Which-Road information, identify a subset
of target vehicles, and refine Which-Lane information for them
using RTK data. Both will use the same OTA message set and
basic onboard processing methods.
Positioning Architecture
The two-stage OTA messaging scheme enables use of this
architecture for a range of implementations. FIGURE 4 shows
www.gpsworld.com
Vehicle Integration | TRANSPORTATION
p FIGURE 5 Which-Road and Which-Lane awareness zones (AZ)
the scalable onboard system, containing two service compo-
nents, GPS and DSRC.
The DSRC transceiver receives OTA messages from all
V2X senders within its communication range, and all senders
are mapped to entities in an AZ (FIGURE 5) in the OTA-local
map module. Position information from heartbeat messages of
other vehicles are decoded and forwarded to the Which-Road
positioning module, where these and data from onboard GPS
estimate the Which-Road relative positions of those vehicles.
The system sends these estimates to V2X applications, which are
coupled with the OTA-local map so that all identified entities
are assigned a tracking priority. The feedback may assign higher
priority to vehicles in the immediate vicinity over vehicles closer
to the AZ boundary. This part of the system is identified as the
Which-Road subsystem.
April 2009 | GPS World 31
TRANSPORTATION | Vehicle Integration
p FIGURE 6 Field test area & V2V and V2I system coverage
The rest of the system contains the
scalable Which-Lane-level positioning
components. This consists of a bank of
RTK vector processors (RTKVP) that
receives raw GPS data from onboard
GPS and the OTA-local map module.
The bank could contain one or more
RTK V Ps, depending on resources
available, and each could be dedicated
to estimate a precise relative vector
between the host vehicle and any other
communicator. For a vehicle RTK data
stream, the precise vector and host
location are used to estimate the precise
relative orientation of the sender vehicle
in the Which-Lane positioning module.
The target priority established in the
OTA-local map and V2X application
feedback loop drives selection of the
vehicles to be precisely positioned using
the Which-Lane part of the system.
Whenever an infrastructure RTK feed is
available as identified by the OTA-local
map, that feed is directed to a RTKVP,
and the RTK relative positioning module
uses the resulting precise vector and
reference information to estimate the
precise position of the host.
In summary, the basic onboard system
configuration enables Which-Road-level
positioning of all communicating vehi-
cles with minimal processing resources.
It allows identification of a subset of ve-
hicles that are of greater interest to the
q FIGURE 7a Northbound drive scenario
32 GPS World | April 2009
applications. Depending on the onboard
resources available, the host system may
refine the relative position of selected tar-
get vehicles using the Which-Lane subsys-
tem. A simplified implementation of the
architecture may be designed without a
Which-Lane subsystem and may enable
only Which-Road applications.
The Which-Road AZ can be estab-
lished with minimal processing require-
ments, using only the Which-Road
subsystem. It also enables target prioriti-
zation. With the full implementation, the
host vehicle is able to establish two AZs
as shown in Figure 5. If an infrastruc-
ture element is within the Which-Road
AZ (which is also the communication
range), the host vehicle absolute position
accuracy can be improved to Which-Lane
level by prioritizing an RTKVP to process
a relative vector between the host and the
selected sender. Otherwise, the host ve-
hicle absolute accuracy could be limited
to Which-Road level depending on the
GPS service quality and availability of
augmentation data.
System at Work
The following tests use data captured
from two vehicles, both equipped with
the basic onboard architecture shown in
Figure 4.
Onboard Configuration. The vehicles
were equipped with a primary GPS con-
figuration using a production-like low-
cost GPS receiver. These L1-only receivers
with 12 or more tracking channels, using
active-patch antennas, are identified as
type A.
A secondary GPS configuration also
went onboard both vehicles, including
a survey-grade, dual-frequency GPS re-
ceiver. These devices are capable of track-
ing L1 C/A and L2 P(Y) signals and have
12 or more channels dedicated to each sig-
nal. They use survey-grade active anten-
nas with advanced multipath mitigation
capabilities and are identified as receiver
type B hereafter.
Type B receivers primarily gathered data
for reference trajectory generation using
post-processing tools. The L1 portion of
this receiver was also used as a second
L1-only GPS device to illustrate the func-
tionality of the architecture when vehicles
with different receiver types operate. Both
receiver types were configured to operate
in standalone GPS mode for the test.
Onboard positioning systems in both
vehicles logged their own GPS status as
well as the output from both Which-Road
and Which-Lane subsystems.
Test Area. The test vehicles were
driven in a North-South loop on a sub-
urban highway with three to four lanes
per direction ( FIGURE 6). The vehicles
stayed within 300 meters of each other
at all times to maintain real-time V2V
communication. The roadway ran mostly
straight with considerable infrastructure
of up to three-story buildings on either
side. The approximately four-mile drive
in one direction lasted about 7–8 minutes.
The test area included three simulated sig-
nalized intersections equipped with V2I
capability. Figure 6 shows coverage of V2I
systems at each intersection and the com-
www.gpsworld.com
 Vehicle Integration | TRANSPORTATION

munication ranges of both V2V-equipped The across-distance plot (bottom) ity and positioning mode differences. In
vehicles (not to scale). shows the estimate available from the time between host and target system
System performance results from a full Which-Road and Which-Lane subsys- switch to RTK mode at intersection A
north-south loop are shown under two tems. Comparing with the reference, the (time 105 seconds), the Which-Road es-
scenarios: when both vehicles use the same Which-Lane subsystem output agrees with timate is completely driven by the absolute
onboard receiver type (type A), and when the reference solution at sub-meter level error difference of the vehicles. During
one uses type A and the other type B. at all times. However, the Which-Road these periods, errors experienced by one
Northbound Drive, Same Type. In the subsystem shows significant deviations vehicle are almost eliminated by intersec-
first test, both vehicles drove northbound at times, mostly driven by satellite visibil- tion RTK data (in RTK mode) while the
through the entire test area (FIGURE 7a).
The vehicles started in the same lane, with
the host following the target. The target
performed two lane-change maneuvers
during the run while the host remained in
the middle lane. During the lane change,
the target first changed to the left adjacent
lane, back to same lane formation with the
host, then changed to the right adjacent
lane to perform a similar maneuver.
For simplicity, the performance analy-
sis used the distances between vehicles,
in particular the between-vehicle across-
distance, for ease of validation. During a
same-lane formation drive, the between-
vehicle across-distance should be close
to zero, while in adjacent-lane drives, it
should approximately equal a lane width.
FIGURE 7b shows the absolute positioning
accuracy observed in both vehicles (top
plot) and the between-vehicle across-dis-
tance (bottom plot) as estimated by the
Which-Road and Which-Lane subsystems
using receiver type A in both vehicles. Also
shown is the reference across-distance gen-
erated by post-processing dual-frequency
data from the type B receivers.
The absolute positioning accuracy plot
shows three time periods in which the ac-
curacy improves from around 5 meters
RMS to better than 1 meter RMS. These
are the durations when the vehicles were
within the coverage of a V2I broadcast,
thus were receiving RTK data from inter-
sections. At intersection A, the target ve-
hicle position solution switches to RTK a
few seconds ahead of the host and remains
in RTK mode for around 35 seconds until
the vehicles depart from the intersection
coverage area. The absolute accuracy plot
shows that infrastructural support can
furnish Which-Lane accuracy. Outside
V2I coverage, both vehicles show typical
standalone GPS accuracy.
www.gpsworld.com April 2009 | GPS World 33
TRANSPORTATION | Vehicle Integration
p FIGURE 7b Positioning system output, northbound drive, GPS type A in both vehicles
errors in the other remain. Similar spikes
are seen every time the vehicles enter or
exit an equipped infrastructure coverage
area at two different times.
In addition to these mode-difference-
driven errors, the Which-Road estimate
shows several other meter-level errors, pro-
duced by differences in the satellite obser-
vations used by the host and target vehicles.
Around 50 and 280 seconds, around 1-
meter deviation was seen in the Which-
Road estimate due to satellite observation
differences. This illustrates the vulnera-
bility of the Which-Road estimate to sky
visibility difference-driven errors, even
with identical receivers on all vehicles,
and vehicles staying within tens of meters
from each other.
The Which-Lane subsystem, however,
shows approximately centimeter-level
agreement with the reference across-dis-
tance throughout the time shown. As ex-
pected, no errors occur from differences
in satellite visibility between vehicles or
due to infrastructural support. Total vis-
ible satellite count in both vehicles stayed
q FIGURE 8a Southbound drive scenario
34 GPS World | April 2009
above four for the entire test.
Southbound Drive, Same Type. In
FIGURE 8a, the same vehicle pair com-
pleted the southbound test loop. The drive
started with both in a same-lane forma-
tion. After intersection C, both changed
lanes to the left, the target vehicle shortly
followed by the host. Then the target re-
peated two maneuvers, each consisting of
a lane change to the right, several seconds
of adjacent-lane formation driving, and a
change back to same-lane formation.
FIGURE 8b shows the absolute error and
across-distance outputs from the host sys-
tem operating with receiver type A from
both vehicles. In contrast to Figure 7b, the
absolute errors in both vehicles are almost
identical throughout the run, demonstrat-
ing how different receivers of the same
make and model can produce different
accuracy when operating in standalone
mode. The absolute errors in both vehicles
reach sub-meter levels when vehicles are in
V2I coverage, as in the northbound run.
In Figure 8b bottom plot, the across-
distance reported by the Which-Road
subsystem agrees with the reference
solution at sub-meter level for most of
the data shown except for the mode dif-
ference-induced errors. Six such abrupt
Which-Road error increases correspond
to entries and exits of V2I coverage areas.
The one at 70 seconds clearly illustrates
how these can influence application per-
formance. At this time, the target vehicle
changed to the left adjacent lane, closely
followed by the host. As indicated by
the reference and Which-Lane subsys-
tem, the maximum net across-distance
variation during this maneuver is a single
lane width, which reaches the maximum
when the target is in the left adjacent lane
and the host is about to start the change
to the same lane. However, due to the net
effect of the lane change and the mode
difference-induced error which starts
at the peak across-distance separation,
Which-Road subsystem appears to show
a further increase of the across-distance
that eventually reached close to two lane
widths. This could make an application
less reliable or could even cause applica-
tions to malfunction.
Different Receiver Types
FIGURE 9 shows the same absolute accu-
racy/across-distance plots as Figure 7b,
but estimated using the L1 data from
receiver type B in the host and type A
in the target. This setup investigated
the contribution of receiver type differ-
ences on the relative positioning system
performance at both Which-Road and
Which-Lane levels.
The absolute accuracy in Figure 9 de-
picts a clear error characteristic difference
between receiver types. In comparison to
Figure 7b, one receiver shows stepwise
changes (host with type B) while the
www.gpsworld.com
 Vehicle Integration | TRANSPORTATION

between individual vehicle absolute errors

p FIGURE 8b Positioning system output, southbound drive, GPS type A both vehicles
other, using type A, shows gradual error
changes over time. This is directly driven
by receiver hardware/software characteris-
tics and often depends on manufacturers’
proprietary algorithms and hardware.
Although not the most prominent error
source, mode-difference-driven errors sim-
ilar to Figure 7b appear clearly in Figure 9,
Which-Road subsystem across-distance.
In addition, Which-Road subsystem
across-distance shows significant errors,
some of which appear as stepwise errors,
such as the one around 375 seconds, and
others showing comparatively smoother
variations, as from 30 to 50 seconds. A
correlation of this error and the difference
indicate that Which-Road subsystem is
entirely driven by receiver-type incom-
patibilities. Around 180 seconds, a step
change in host accuracy translates into an
across-distance error of around 2 meters.
While this is the most obvious instance of
such an occurrence, several other times in
which receiver difference contributed to
Which-Road-level system errors of several
meters occur in the dataset (230, 250, to
275 seconds). The Which-Lane subsys-
tem, on the other hand, shows good ac-
curacy within tens of centimeters from the
reference solution. In comparison to Fig-
ure 7b, which shows Which-Lane output
when the same receiver type is used in both
vehicles, Which-Lane output in Figure 9
shows almost identical accuracy and avail-
ability performance. This confirms that
receiver types A and B are interchangeable
in a V2X Which-Lane subsystem that uses
RTK techniques, while it may not be the
case in a Which-Road subsystem imple-

5
u-blox 5
Under 1 second Time To First Fix
50 channels, 1 million correlators
SuperSense® -160 dBm sensitivity
Under 50 mW power consumption
Galileo-ready

u-blox’ 5th generation has redefined

the boundaries of GPS performance

GPS for Mobile Terminals

To experience u-blox 5 performance first-hand, come visit us at:

your position is our focus Mobile World Congress 2008, Feb. 11-14, Barcelona, Spain
CeBIT 2008, March 4-9, Hannover, Germany
CTIA 2008, April 1-3, Las Vegas, NV, USA

Single Chips Chipsets Modules www.u-blox.com

www.gpsworld.com April 2009 | GPS World 35

TRANSPORTATION | Vehicle Integration
p FIGURE 9 Positioning system output using receiver types A (target) and B (host)
p FIGURE 10 Positioning system output using receiver types A (target) and B (host)
mentation, as it shares only the position
estimates between vehicles.
FIGURE 10 shows the absolute error and
across-distance output during the south-
bound drive with type A and B receivers.
The absolute error difference in this case is
significant compared to the northbound
drive, reaching a difference as high as
4 meters at 100 seconds. Examination of
the visible constellation showed the host
vehicle drove next to a truck for most
of this period, obstructing almost half
of the host’s visible sky and producing a
substantial horizontal error. However, as
seen from the across-distance plot, most
of this error was translated to the along-
track distance component due to the ve-
hicle heading and its relationship with the
error vectors at the time.
36 GPS World | April 2009
In contrast, at 275 seconds the absolute
error difference of approximately 2 me-
ters is directly translated into an across-
distance error. Towards the end of the run,
starting around 350 seconds, Which-Road
subsystem across-distance error of more
than 2 meters occurs even when the abso-
lute errors are approximately equal. Fur-
ther examination shows that although the
absolute error magnitude is approximately
equal during this time, their orientation
was different. Once these were translated
into across-distance, a net error of around
2 meters in across-distance remained.
Conclusions
Tests show that Which-Lane position-
ing capability is the critical enabler for
a suite of V2V ADAS functions. While
Which-Road implementation is simpler,
several technical challenges remain. Er-
rors will likely stem from sky-visibil-
ity differences, vehicle GPS positioning
modes, and receiver hardware/software
differences. Using GPS measurements
to overcome these limitations brings
significant advantages while greatly
improving the robustness of relative
positioning. Use of both Which-Road
and Which-Lane components and in-
telligent onboard data fusion most ef-
fectively enables V2X positioning while
adhering to real-world constraints.
From a system design perspective,
this architecture and messaging scheme
is scalable and flexible enough to serve
as the core positioning architecture
under a range of vehicle capabilities and
infrastructure support. It could adapt to a
V2V environment with no infrastructural
support and can be switched to enable
infrastructure support if such capability
is available, with minor configuration
changes. It can further extend to depend
solely on infrastructure support if wide
V2I coverage is available. Design also
allows choice of the type of applications
that can be supported. Some prototype
vehicles may support a full application
suite including Which-Lane, while
another prototype class may support
only Which-Road applications. Most
important, none of these cases require any
architectural or OTA messaging changes.
GM has years of experience in V2X
positioning technology development and
currently works with automotive and
industry partners on standardization and
basic system development efforts.
Acknowledgments
Thanks to GM R&D’s ADAS team,
especially to Hariharan Krishnan and
Donald K. Grimm, and to Cem U
Sraydar, Manager, Wireless Systems and
Technology Group.
CHAMINDA BASNAYAKE is a senior research engi-
neer at General Motors R&D and Planning,
and GNSS technology expert for GM OnStar.
He has a Ph.D. in geomatics engineering
from the University of Calgary, Canada. He is
also the GNSS technical lead of an automo-
tive consortium working on V2X R&D under
a corporative agreement with U.S. DOT.
www.gpsworld.com

Worst-Case Scenario
SBAS
Interference
Some applications using data from fixed or low relative
dynamic WAAS or other SBAS receivers may encounter
a tracking error due to C/A code cross-correlation.
L. Lestarquit, J-L. Issler, O. Nouvel, M. Sihrener, G. Lamontagne, J.C. Guay, R. Landry, O. Julien, C. Macabiau, M. Nouvel-Malicorne
G
PS C/A codes present cross-correlation peaks that
can potentially cause false acquisitions. They can also
create tracking errors and C/N0 degradation, not in
traditional GPS tracking situations, but for applications where
signals are received with a low Doppler difference.
Several applications present a low relative Doppler, thus caus-
ing more frequent Doppler collision and this type of interference.
One occurs with the use of ranging signals from geostationary
satellites (for example, satellite-based augmentation systems, or
SBAS) when two or more of these satellites are in line-of-sight of
a fixed- or low-dynamic receiver such as a control station, refer-
ence station for differential positioning, or any ground user with
little or no motion. The tracking error increases strongly with
more geostationary satellites — and multiple SBAS have been
and continue to be developed with overlapping interoperative
regions. Indeed, it is possible that all the signals will interfere at
the same time, or sequentially, one after another. The SBAS C/A
code interference could also be seen for mobile receivers such as
airplanes, moving along a zero-relative Doppler line between two
SBAS satellite longitudes.
Pseudolite use can also induce low dynamic signals, in applica-
tions involving a pseudolite and a receiver onboard a geostationary
satellite, and pseudolites and fixed- or low-dynamic receivers.
The interference mechanism is physically the same as for
multipath error. Actually, multipath can be viewed as C/A code
self-interference: the cross-correlation peak of the reflected signal
contributes to the correlation function and causes an offset of the
zero-crossing of the discrimination function, known as the track-
ing error. In C/A code interference, the cross-correlation peak of
the interfering code creates the same effect. Normalized cross-
correlation peak values are either Ⳮ63/1023 or ⳮ65/1023. The
error can reach 18 meters for a half-chip spacing correlator when
two quasi-stationary C/A codes are received with the same power
level. If the cross-correlation function of desired and interfering
codes has a secondary peak within the receiver’s chip spacing of
the relative code offset, there is potential for a tracking error.
A necessary condition for such interference is a relative Dop-
www.gpsworld.com
Signal Processing | SYSTEM DESIGN & TEST
pler between the two interfering signals lower than the receiver
code loop bandwidth, typically lower than 1 Hz. The signed
amplitude of the interfering peak is a function of the relative
phase, and if the relative phase moves too much during the code-
loop integration period, the interfering peak (and the tracking
error) will be filtered out. We refer to this condition as a Doppler
collision, or as having a quasi-stationary code, because from the
receiver point of view, the received codes will not move relative
to one another. This means that the cross-correlation peak of
the interfering signal will not move relative to the main auto-
correlation peak of the desired signal, and the interference will
result in a lasting bias.
For common GPS applications, a Doppler collision smaller
than 1 Hz is quite a rare phenomenon. Doppler due to satellite
motion is high (from ⳮ4.5 kHz to 4.5 kHz for standard use),
giving a low probability of such a Doppler collision; even if it
happens, it will last a few seconds. For common GPS applications,
C/A code interference rarely leads to a noticeable tracking error.
However, with the development of GNSS systems using geo-
stationary satellites such as EGNOS, WAAS, GAGAN, MSAS,
BEIDOU-1, or IRNSS, the dynamics are much lower, in the
range of a few meters/second, corresponding to a few 10s of Hz
at L1 frequency. Doppler collisions will occur twice a day and last
several minutes. Thus, interference due to PN-code code cross-
correlation will occur, and can cause false acquisition, C/N0 deg-
radation, and tracking errors. Our results using data collected by
WAAS receivers in Canada show the impact of several factors,
including the navigation message, on the tracking error.
The current EGNOS inhibits the GEO ranging function, and
therefore pseudorange measurement errors made on the GEO
signal have no impact today on EGNOS performance. Moreover,
in more than 10 years of operation, including the EGNOS test
bed, no message loss due to the C/A code interference has been
observed or reported on EGNOS. Our work here intends to:
䡲 precisely characterize these interference tracking errors,
䡲 identify the factors driving these errors, and
䡲 propose mitigation techniques to reduce them.
April 2009 | GPS World 37
SYSTEM DESIGN & TEST | Signal Processing

MULTIPATH
Retard relatif: -0.5, Puissance relative: 0
Retard relatif: -0.5, Puissance relative: 5
40 Retard relatif: -0.5, Puissance relative: 10
Retard relatif: 0.5, Puissance relative: 0
Retard relatif: 0.5, Puissance relative: 5
30 Retard relatif: 0.5, Puissance relative: 10
C/A CODE CROSS CORRELATION INTERFERENCE

Erreur de position (m)

SIGNAL A - DESIRED 20
SIGNAL B - JAMMER
TRACKING 10
CHANNEL 1
SIGNAL B - DESIRED 0
SIGNAL A - JAMMER
-10
TRACKING
CHANNEL 2
-20
Å FIGURE 1 C/A code cross-correlation interference
-30
0 π/2 π 3π/2 2π
Déphasage: 0
Puisance relative: 0 db Déphasage (rad)
10 Déphasage: pi/3
Déphasage: pi/2 Å FIGURE 3 Theoretical tracking errors for carrier-phase measurement
8 Déphasage: pi
6 the code loop (~1 Hz) occurs only twice a day per pair of SBAS
Erreur de position (m)

4 satellites, and lasts from a few minutes up to one hour or more in

2 the maximum interference zones describes hereafter.
0

-2 Tracking-Error Theory
-4 In the presence of an interfering signal, the receiver delay lock
-6 loop not only tracks the desired signal delay, but also the delay of
-8 the sum of the desired signal (index k) and all interfering signals
-10
-1.5 -1 -0.5 0 0.5 1 1.5
(index i ⫽ 1 to N).
Retard relatif (m) The received signal can thus be written as:
Å FIGURE 2 Theoretical tracking errors for pseudorange measurement

Applications Where : PRN code sequence of the ith satellite

FIGURE 1 shows the similarities between multipath (interference : data from the ith satellite
of the delayed signal on the direct signal) and C/A code cross- : pulsation of the reference signal, 2 ⫻ ␲ ⫻ frequency
correlation interference (interference of a cross-correlation peak , and : power, propagation delay, and phase of
on the desired signal main peak). the ith signal with respect to the desired one.
Theoretical error is 9 meters for normal cross-correlation With an early-minus-late power code discrimination function
configurations (for ½-chip spacing correlator and equal power and in the presence of low Doppler, the error function induced by
signals) when two received C/A codes have very small relative a C/A code interference can be written as:
Doppler and the same power. The interfering signal is filtered
at several levels:
䡲 Pre-detection integration will filter the interfering signal if
its relative Doppler is larger than the pre-detection band-
width (50 Hz).
䡲 Code-loop filter will filter the IF signal if its relative Dop-
pler is larger than the DLL bandwidth (<1 Hz).
There is potentially a tracking error only if the relative Doppler
between the received signals is lower than the DLL bandwidth
(typically < 1 Hz).
Satellite Motion. Geostationary satellites’ motion is supposed
to be fixed relative to the Earth. Therefore, signals from SBAS
geostationary satellites would always be in a Doppler collision
with each other when received by a fixed receiver. In reality, geo-
stationary satellites move slowly within a station-keeping window,
mostly along the N-S and the E-W axis, within a 24-hour period.
This generates relative Doppler with a harmonic pattern and an
amplitude of several 10s of Hz, so that a Doppler collision inside
38 GPS World | April 2009 www.gpsworld.com
 Signal Processing | SYSTEM DESIGN & TEST

Where: WAAS message correlation zone.

: cross-correlation between desired
signal k and ith signal
: delay lock loop bandwidth
: C/A code period
: C/A code length.
From this expression we observe that
the tracking error induced by C/A code
interference will depend on different
parameters:
the relative code delay:
the power factor:
the relative carrier phase:
Such a zone exists for each pair of SBAS
satellite signals using C/A code. The lon-
gitude range of these zones will be larger
in the future, since more and more geo-
stationary satellites exhibit very accurate
station-keeping thanks to a stable north-
south ionic propulsion subsystem. These
zones move along the east-west axis, since
Å FIGURE 4 Assumption
the geo-satellites perform E-W or W-E ma-
neuvers from time to time. The synchroni-
zation of WAAS signals at emission made
our zero-Doppler zone prediction easier.
The measurement campaign had a 3-

the relative Doppler

the cross-correlation function:
and the message data :
FIGURE 2 shows theoretical tracking
errors for the pseudorange tracking error
(in meters) versus the relative code delay
(in chip units) for different phase shifts
between the two received carriers.
O M N I S T A R H I G H P E R F O R M A N C E TM D G P S :
FIGURE 3 shows theoretical tracking er-
rors for the carrier-phase tracking error (in
meters) versus the relative phase delay (in BRINGING DECIMETER
radians) for different phase shift and dif-
ferent relative received powers between the
ACCURACY WITHIN REACH
two received carriers.
Assumptions for these theoretical
results are: no relative Doppler, ½-chip
correlator spacing, infinite bandwidth,
no data message, and a cross-correlation
function as in FIGURE 4.
Introducing OmniSTAR High Performance (HP), our revolutionary new DGPS
system that redefines the standard for differential correction. OmniSTAR has
Error Observations
developed cost-effective, reliable real-time corrections enabling high-accuracy
We characterized C/A code interference
applications over entire continents without deploying local base stations.
with three complementary methods:
Our innovative HP signal provides five times the accuracy of standard
䡲 computing the 3rd-order derivative of
DGPS for horizontal accuracy of ~ +/- 10-15 cms. HP means a wide-area
the pseudorange, to detect anomalies
䡲 analyzing the measurement residual solution offering previously unobtainable levels of accuracy, opening the door
errors obtained by interpolation (over for high-precision uses such as auto steering, guidance and automation
a sliding two-hour window) and of agri-vehicles, increased hit probability for GIS/utilities, rough grading
extraction of the range error for construction applications, survey projects, oil exploration, aerial
䡲 fine analysis of relative Doppler and photogrammetry, and more. We just raised the bar on accuracy and reliability,
carrier phase; Gold code sequences but it’s way within your reach. OmniSTAR HP. Others promise—we deliver.
cross-correlation function; SBAS
navigation messages correlation.
All the data are measurements taken
© 2005, OmniSTAR

June 18–22, 2008, in British Columbia,

Canada, at a location where our geometri-
Contact us for more information ®
cal analysis predicted a zero relative Dop- www.omnistar.com Expect high performance
pler between two WAAS satellites. The
green band in FIGURE 5 depicts this highest
www.gpsworld.com April 2009 | GPS World 39
SYSTEM DESIGN & TEST |
Å FIGURE 5 Highest WAAS message correlation zone
Å FIGURE 6 Observation site near Kamloops, British Columbia, Canada
hour trial recording period, and a 48-hour
continuous recording period.
FIGURE 6 shows the location, near Wal-
hachin, 45 kilometers from Kamloops.
FIGURE 7 shows a very long Doppler
collision, lasting 1 hour, 45 minutes. The
results in FIGURE 8 are related to the cross-
Å FIGURE 7 SBAS Doppler evolution
40 GPS World | April 2009
Signal Processing
correlation function.
The results in FIGURE 9 are related to
the cross-correlation error residuals. PRN
135 is received with 2.5 dB more power
than PRN 138. Errors seem to comply
with theory.
The results in FIGURE 10 are related to
Å FIGURE 8 Relative Doppler between PRN 135 and 138
the data message correlations. Change in
data message correlation level has been no-
ticed from January to June 2008.
Analysis
The C/N0 measurements are clearly af-
fected by the cross-correlation interfer-
ence. Pairs of C/N0 degradation periods
are clearly visible during each Doppler
collision on June 19. These pairs are
explained by the slow shift of the cross-
correlation function. We measured deg-
radations of about 0.5 dB.
These pairs can be observed on the 3rd-
order derivative of the pseudorange. These
C/N0 degradations must be taken into ac-
count in the WAAS link budgets, and in
the computation of the L1 power increase
of SBAS modernization regarding L1. Fig-
ure 10 gives an example of such a 3rd-order
pseudorange derivative for PRN 135, for
the same period as the PRN 135 pseudor-
ange measurement in Figure 9.
Conclusions
This study set out to analyze in detail an
error affecting positioning performance
for users and applications using GEO
ranging measurements as offered today
by some SBAS. Such is the case for the
U.S. WAAS and the Japanese MSAS.
It is not the case for the European
EGNOS, as EGNOS users are not
authorized to use these GEO pseudo-
range measurements in the system’s cur-
rent version.
The positioning error mechanism ana-
lyzed here, due to interferences between
SBAS codes during Doppler collisions,
is similar to the case of multipath errors.
www.gpsworld.com
 Signal Processing | SYSTEM DESIGN & TEST

Å FIGURE 9 PRN 135 and 138 cross-correlation function Å FIGURE 10 PRN 135 and 138 relative delay
were not observed at the end of our mea-
surement campaign, because one of the
two observed WAAS satellite made an
east-west maneuver during that period,
this having displaced the maximum in-
terference zone.
If n is the number of SBAS satellites in
a given region of the world, the number
of maximal interference zones covering
the region is n!/[2!⫻(n-2)!]. Four SBAS
over a region will produce six maximal in-
Å FIGURE 11 PRN 135 error w/o smoothing Å FIGURE 12 PRN 138 error w/o smoothing terference zones. Five SBAS satellites will
The most affected applications are those curate ranging. At a minimum, it would produce 10 maximal interference zones in
of fixed SBAS users, or ones provided with be necessary in case of GEO ranging to the region, and so on. These zones are not
low dynamics relative to the Doppler. detect the errors, and either to eliminate stable since the geostationary SBAS space-
Results show a possibility of mitigating all measurements made during Doppler craft maneuver sporadically.
the interference by applying a correction collision from the navigation solution, or The SBAS C/N0 degradation at L1 due
algorithm using the method suggested to adequately bound the associated errors to quasi-stationary C/A codes is due to a
by CNES: for instance, in all the user in the integrity computation. variation of the C component, rather than
WAAS receivers supposed to perform ac- The worst cases of Doppler collision a variation of the N0 component. We rec-

𰀲𰁁𰁋𰁏𰁎𰀀𰀍𰀀𰁌𰁅𰁁𰁄𰁉𰁎𰁇𰀀𰁔𰁈𰁅𰀀𰁗𰁁𰁙𰀀𰁗𰁉𰁔𰁈𰀀𰁃𰁒𰁙𰁓𰁔𰁁𰁌𰀀𰀀
𰀆𰀀𰀴𰀣𰀸𰀯𰀀𰁔𰁅𰁃𰁈𰁎𰁏𰁌𰁏𰁇𰁙𰀀𰁆𰁏𰁒𰀀𰁏𰁖𰁅𰁒𰀀𰀒𰀐𰀀𰁙𰁅𰁁𰁒𰁓𰀀
𰀯𰁕𰁔𰁓𰁔𰁁𰁎𰁄𰁉𰁎𰁇𰀀𰁐𰁅𰁒𰁆𰁏𰁒𰁍𰁁𰁎𰁃𰁅𰀀𰁉𰁎𰀀𰀧𰀰𰀳𰀌𰀀𰁃𰁏𰁍𰁍𰁕𰁎𰁉𰁃𰁁𰁔𰁉𰁏𰁎𰁓𰀌𰀀𰀀
𰁁𰁕𰁔𰁏𰁍𰁏𰁔𰁉𰁖𰁅𰀌𰀀𰁁𰁅𰁒𰁏𰁓𰁐𰁁𰁃𰁅𰀀𰁁𰁎𰁄𰀀𰁍𰁉𰁌𰁉𰁔𰁁𰁒𰁙𰀀𰁁𰁐𰁐𰁌𰁉𰁃𰁁𰁔𰁉𰁏𰁎𰁓𰀀𰁁𰁒𰁏𰁕𰁎𰁄𰀀𰁔𰁈𰁅𰀀𰁇𰁌𰁏𰁂𰁅𰀎
𰀳𰁕𰁐𰁅𰁒𰁉𰁏𰁒𰀀𰁐𰁈𰁁𰁓𰁅𰀀𰁎𰁏𰁉𰁓𰁅𰀀𰁐𰁅𰁒𰁆𰁏𰁒𰁍𰁁𰁎𰁃𰁅
𰀦𰁒𰁅𰁑𰁕𰁅𰁎𰁃𰁙𰀀𰁇𰁒𰁁𰁄𰁉𰁅𰁎𰁔𰀀𰀜𰀐𰀎𰀐𰀕𰁐𰁐𰁍𰀏𰂪𰀣
𰀋𰀏𰀍𰀀𰀐𰀎𰀕𰁐𰁐𰁍𰀀𰁆𰁒𰁅𰁑𰁕𰁅𰁎𰁃𰁙𰀀𰁓𰁔𰁁𰁂𰁉𰁌𰁉𰁔𰁙𰀀𰁏𰁖𰁅𰁒𰀀𰁗𰁉𰁄𰁅𰀀𰁔𰁅𰁍𰁐𰁅𰁒𰁁𰁔𰁕𰁒𰁅𰀀𰁒𰁁𰁎𰁇𰁅𰁓

𰁏𰁒𰁆𰁎𰀃𰁒𰁑𰀃𰁚𰁌𰁗𰁋𰀃𰀵𰁄𰁎𰁒𰁑

𰁉𰁌𰁑𰁇𰀃𰁗𰁋𰁈𰀃𰁅𰁈𰁖𰁗𰀃𰁓𰁕𰁒𰁇𰁘𰁆𰁗𰀃𰁗𰁒𰀃𰁖𰁘𰁌𰁗𰀃𰁜𰁒𰁘𰁕𰀃𰁑𰁈𰁈𰁇𰁖𰀃𰁄𰁗𰀝𰀃𰁚𰁚𰁚𰀑𰁕 𰁄 𰁎 𰁒𰁑𰀑𰁆𰁒 𰁐

www.gpsworld.com April 2009 | GPS World 41

SYSTEM DESIGN & TEST | Signal Processing

Å FIGURE 13 Nav message cross-correlation, PRN 135 and 138 Å FIGURE 14 Nav message cross-correlation, PRN 135 and 138

ommend that this effect be taken into account in detailed radio

frequency compatibility studies between GNSS on L1.

LAURENT LESTARQUIT, JEAN-LUC ISSLER and MARIE NOUVEL-MALICORNE work at

the French Space Agency; OLIVIER NOUVEL and MATTHIEU SIHRENER at M3
Systems, France; GUILLAUME LAMONTAGNE, J.C. GUAY, and RENÉ LANDRY, JR. at
the École de Technologie Supérieure, Canada; and OLIVIER JULIEN and
CHRISTOPHE MACABIAU at the Ecole Nationale de l’Aviation Civile, France.

Manufacturers
Tests used a NovAtel DL-4/OEM4 receiver (www.novatel.com ).
Å FIGURE 15 C/N0 of WAAS PRN 135 and 138

COME TO
ANNAPOLIS
GNSS Education
Annapolis, Maryland
June 15-19, 2009
GPS Operation for Engineers
and Technical Professionals
Course 356 (5 Days)

GPS Fundamentals &

Enhancements
Course 122 (2 Days)

DGPS, GPS Signals & Processing

Course 356B (3 Days)
Applied Kalman Filtering
Course 447 (4.5 Days)

NavtechGPS
www.NavtechGPS.com
800.628.0885 or 703.256.8900
courses@navtechgps.com
42 GPS World | April 2009 www.gpsworld.com
 Algorithms & Methods | INNOVATION

Precise Point Positioning

A Powerful Technique with a Promising Future
Sunil Bisnath and Yang Gao

T
he main goal of this article is to
describe the current performance
of what has become known as the
precise point positioning (PPP) technique,
and to discuss the future potential of
the technique, along with its technical
limitations. We begin with a review of the
current state of PPP, covering performance
and usage. We then discuss current
technical limitations of the approach,

MORE THAN 10 YEARS AGO in an Innovation column, I wrote, “Although RTK

is the latest word, or should we say acronym, in GPS positioning, it will not be
the last. Scientists and engineers will continue to invent faster, more accurate,
more convenient, and more reliable ways to use GPS in navigation, surveying,
and a host of other areas, some of which we haven’t even dreamt of yet.”
In the intervening decade, RTK — or real-time kinematic — positioning
has become an industry standard procedure in surveying, machine control,
and other high-precision applications. RTK makes use of carrier-phase and
pseudorange measurements recorded at
a (usually) fixed reference location with
known coordinates and transmitted in real
time to a user’s rover receiver using a radio
link of some kind. The rover processes the
double differences of observations between
satellites and receivers to determine its
INNOVATION INSIGHTS coordinates with better than 10-centimeter
with Richard Langley accuracy. It can do this successfully if it can
resolve the integer ambiguities in the carrier-
Is there a viable phase measurements. Ambiguities are the
bane of carrier-phase positioning. They must
alternative to RTK? be resolved to turn carrier-phase measure-
ments into unbiased range measurements.
In RTK positioning, the ability to resolve ambiguities is determined by
many factors, such as the distance between the reference station and the
rover and atmospheric effects. RTK is a much more efficient technique than
the earlier developed (but sometimes still used) post-processing surveying
techniques. However, it does require an investment in reference station in-
frastructure or the purchase of commercial RTK services.
Is there a viable alternative to RTK? In this month’s column, we take a look
at the technique of precise point positioning (PPP). Like RTK, PPP makes use
of ambiguous carrier-phase measurements but only from the user’s receiver.
Rather than measurements from a reference receiver, it needs ultra-precise
(and accurate) satellite orbit and clock information such as that provided by
the International GNSS Service. Currently, there are issues with how long
solutions take to converge and the difficulty in resolving the ambiguities, for
example, but research is targeting these and other practical issues. How
close is PPP to prime time? Read on.
“Innovation” is a regular column that features discussions about recent advances in GPS technology and
its applications as well as the fundamentals of GPS positioning. The column is coordinated by Richard
Langley of the Department of Geodesy and Geomatics Engineering at the University of New Brunswick, who
welcomes your comments and topic ideas. To contact him, see the “Contributing Editors” section on page 4.

www.gpsworld.com April 2009 | GPS World 43

INNOVATION | Algorithms & Methods
including solution convergence period, accuracy, and integrity
of solutions. The next section considers potential improvements
upon the current approach, in terms of integer ambiguity
resolution; integration with other data, for example, from
real-time kinematic (RTK) solutions or an inertial navigation
system (INS); and the use of other external modeling data, such
as atmospheric refraction models. Equally important are PPP-
infrastructure challenges, including the availability of precise
satellite orbits and clock offsets, precise orbit and clock prediction,
real-time dissemination of predicted orbits and clocks, and
reference frame realizations. Given the upcoming great changes
due to GPS modernization and the development of other global
navigation satellite systems (GNSS), we would be remiss not to
speculate on the potential significant positive impacts of these
added signals on future PPP performance. Finally, we end with a
rather provocative discussion of the potential of PPP to perform
in a similar manner as the RTK technique.
Current Status
This section is designed to summarize current PPP performance
using a number of metrics, and to set the technique’s impact
within the context of the wider field of positioning and naviga-
tion. What we mean by PPP is the state-space solution to the pro-
cessing of pseudorange and carrier-phase measurements from a
single GNSS receiver, utilizing satellite constellation precise orbits
and clock offsets determined by separate means. Typically, a dual-
frequency GNSS receiver is used with dual-frequency code and
phase measurements linearly combined to remove the first-order
effect of ionospheric refraction. The real-valued carrier-phase am-
biguity terms are estimated from the measurement model. The
tropospheric refraction is also estimated, along with the receiver
position and ambiguity parameters from the measurements. PPP
using a single-frequency GNSS receiver has also been investigated
with great promise for certain applications. However, we will
not discuss these further in this article; see Further Reading for
publications reporting developments in single-frequency PPP
and other advances in the technique. To achieve the best position
accuracy possible from PPP, effects such as carrier-phase wind-up,
transmitter-antenna phase offset, solid Earth tides, and the con-
tribution of ocean tide loading must be corrected using models.
Residual terms such as receiver noise and multipath are generally
ignored or minimally handled using stochastic procedures.
A unique aspect of PPP is that it is an area of research being
actively pursued by academia, government, and industry, in con-
cert and individually. As is typical, early development occurred in
research settings for scientific goals. Governments, as service pro-
viders, have in some cases engaged in providing PPP services to the
public, given the socioeconomic benefits. Industry has embraced
and advanced the technology to better serve its clients. The results
are: 1) rapid development and use of PPP in a variety of applica-
tion areas, and 2) significant overlap between the three sectors in
terms of research and development, and service models. The latter
point will be discussed further in the infrastructure section.
44 GPS World | April 2009
1.0
3D root-sum-square error (meters)
1/Jul/04
0.8 2/Jul/04
3/Jul/04
0.6 4/Jul/04
5/Jul/04
6/Jul/04
0.4 7/Jul/04
0.2
0
0:00
0:15
0:30
0:45
1:00
1:15
1:30
1:45
2:00
2:15
2:30
2:45
3:00
Time from session start (hours:minutes)
▲ FIGURE 1 St. John’s (STJO) PPP solution convergence on seven
sequential days.
Performance Specifications
The standard metrics used here to describe the performance of
conventional PPP services are: accuracy, precision, convergence
period, availability, and integrity. With PPP solutions showing
very little in the way of biases — typically a few centimeters at
the most — there is very little difference between the accuracy
and precision metrics. In terms of north, east, and up component
accuracies at the 1-sigma level, PPP is able to provide few-
centimeter-level results in static mode and decimeter-level results
in kinematic mode; both could be achieved in either post-mission
analyses or in real time.
The convergence period, namely the length of time required
from a cold start to a decimeter-level position solution, is typi-
cally about 30 minutes under normal conditions and will be sig-
nificantly longer before the position solution can converge to the
few-centimeter level, if at all. This period is determined by the
measurement strength of the observables for a GPS-only solution,
the geometry of the problem, and the redundancy available for
the estimation problem. Initial solutions rely almost exclusively
on noisy pseudorange measurements, the uncertainties of which
are magnified via the ionosphere-free linear combination. The
availability of solutions is usually high, given that application
areas for this technique are open sky, continuously unobstructed
environments. Otherwise PPP would typically not be used (at
its current level of development). Finally, the availability of in-
tegrity measures for the PPP solution is considered. Aside from
filter covariance estimates, quantitative quality measures of the
obtained results are limited. For example, knowledge of biases
in corrections such as precise satellite orbit and clock products,
the potential for biases in estimated coordinates, and measure-
ment outliers are typically not considered (that is, not rigorously
specified and accounted for) in solutions.
Use and Applications
PPP can be used for the processing of static and kinematic data,
both in real time, if the dissemination mechanisms are in place to
construct, transmit, receive and process precise satellite orbit and
clock products, and in post-processing mode. The caveat for all
such usage is that there needs to be uninterrupted GNSS signal
availability, as loss of tracking lock on a minimum number of sat-
ellites requires processing filter re-initialization, resulting in tens
of minutes of greater than decimeter-resolution positioning, until
www.gpsworld.com
 Algorithms & Methods | INNOVATION

1.0

3D root-sum-square error (meters)

0.9 ALBH(L3S)
NRC1(L3S)
filter re-convergence. This constraint severely limits the utility of 0.8
YELL(L3S)
0.7 STJO(L3S)
PPP, insofar as it can only be robustly (that is, successfully) used
0.6
in environments with continuous open sky coverage. 0.5
One of the first uses of a prototype PPP approach was for 0.4
the rapid post-processing of static geodetic data for establishing 0.3
0.2
and updating reference-station coordinates or for crustal-
0.1
deformation monitoring. Other scientific uses range from precise 0
orbit determination of low-Earth-orbiting satellites for gravity 0:00 1:00 2:00 3:00 4:00 5:00 6:00
Time from session start (hours:minutes)
field recovery to ocean buoy positioning for tsunami detection.
▲ FIGURE 2 PPP solution convergence for stations Albert Head
The main commercial applications of PPP have been in the (ALBH); National Research Council, Ottawa (NRC1), Yellowknife
agricultural industry for precision farming; in marine applications (YELL), and St. John’s (STJO). L3S = ionosphere-free, static.
for sensor positioning in support of seafloor mapping and marine
construction, for example; and in airborne mapping. effectiveness, its more widespread use is limited by convergence
Further growth in current active areas is underway, and the period, accuracy, and integrity issues.
technique is making inroads into other application areas such as Convergence Period. PPP requires a long initialization period
atmosphere remote sensing, precise time transfer, land surveying, for phase ambiguities to converge to near constant values and for
construction, and military uses. Fundamentally, PPP is a viable the solution to reach its optimal precision, taking full advantage
option wherever precise positioning and navigation is required in of the precise but ambiguous carrier-phase observations. PPP
isolated locations or expansive areas and reference station infra- convergence depends on a number of factors such as the number
structure is not available, or very costly to temporarily erect. and geometry of visible satellites, user environment and dynamics,
observation quality, and sampling rate. As these different factors
Technical Limitations interplay, the period of time required for the solution to reach a
Although the PPP approach presents definite advantages for pre-defined precision level will vary.
many applications in terms of operational flexibility and cost- Shown in FIGURE 1 are the convergence times with respect

GPS Signal Distribution

even in adverse environments!

High Performance Amplifiers, Splitters, Combiners,

Agriculture Aviation Military Public Safety
Repeater Kits, Antennas, Fiber Optic Links and More!

𰀧𰀰𰀳𰀏𰀳𰀡𰀴𰀣𰀯𰀭𰀀 • Custom solutions with prompt delivery

𰀡𰀮𰀴𰀥𰀮𰀮𰀡𰀀𰀳𰀯𰀬𰀵𰀴𰀩𰀯𰀮𰀳 • Expert GPS and RF technical support
High performance antennas for precise operation, • Regulatory compliant designs
maximum durability, and ease of installation for
applications in:
𰁳𰀀𰀮𰁅𰁔𰁗𰁏𰁒𰁋𰀀𰀴𰁉𰁍𰁉𰁎𰁇 𰁳𰀀𰀡𰁖𰁉𰁁𰁔𰁉𰁏𰁎
𰁳𰀀𰀶𰁅𰁈𰁉𰁃𰁌𰁅𰀏𰀡𰁓𰁓𰁅𰁔𰀀𰀴𰁒𰁁𰁃𰁋𰁉𰁎𰁇 𰁳𰀀𰀰𰁒𰁅𰁃𰁉𰁓𰁉𰁏𰁎𰀀𰀭𰁅𰁁𰁓𰁕𰁒𰁅𰁍𰁅𰁎𰁔
𰁳𰀀𰀰𰁕𰁂𰁌𰁉𰁃𰀀𰀳𰁁𰁆𰁅𰁔𰁙 𰁮𰀀𰀡𰁇𰁒𰁉𰁃𰁕𰁌𰁔𰁕𰁒𰁅
𰁳𰀀𰀭𰁉𰁌𰁉𰁔𰁁𰁒𰁙 𰁮𰀀𰀣𰁏𰁎𰁓𰁔𰁒𰁕𰁃𰁔𰁉𰁏𰁎
𰁳𰀀𰀳𰁁𰁔𰁅𰁌𰁌𰁉𰁔𰁅𰀀𰀣𰁏𰁍𰁍𰁕𰁎𰁉𰁃𰁁𰁔𰁉𰁏𰁎𰁓

Providing friendly service and personalized

technical support

A LEADING PROVIDER OF HIGH PERFORMANCE GPS/SATCOM ANTENNAS

phone 630.372.6800
toll-free 800.323.9122
website antenna.pctel.com
email sales.antenna@pctel.com
Toll Free Phone: 866-289-4777 • www.gpssource.com

www.gpsworld.com April 2009 | GPS World 45

INNOVATION | Algorithms & Methods

Latitude Longitude Height

1.0
0.8
0.6
in FIGURE 4 are the positioning results of a kinematic dataset
Position error (meters)

0.4 acquired from an airborne platform flying at an altitude of

0.2 approximately 250 meters above the ground at 50 knots. The
0.0 precise orbit and clock corrections are again from JPL’s real-time
-0.2 orbit and clock product. The short baseline, double-differenced,
-0.4 ambiguity-fixed position solutions were used as ground-truth,
-0.6
and the positioning accuracy statistics are given in TABLE 2.
-0.8
-1.0
The results indicate that it takes about 20 to 30 minutes for the
259200 273600 288000 302400 316800 331200 345600 positioning solution to converge to the decimeter level.
GPS Time (seconds)
▲ FIGURE 3 Static positioning using a dataset from Algonquin Park
In addition to further improvement of the precise orbit and
clock products, the ability to exploit the integer property of phase
to different position accuracies from processing data from an ambiguities can further improve the obtainable position accuracy
International GNSS Service (IGS) tracking station over seven of PPP. Minor error sources, including initial satellite and
consecutive days. The station data was collected at 30-second receiver phase biases, would need to be estimated and removed
intervals, and each 24-hour session was processed using IGS in the measurement model, as they cannot be eliminated in
precise orbits and satellite clock values at 5-minute epochs. undifferenced processing. We will elaborate on these effects later
The position error was computed every 15 minutes. The results in the article.
show a high degree of day-to-day variability, despite the similar Integrity. Integrity monitoring is an essential component of
satellite geometry one would expect given a common daily session any positioning or navigation system. In PPP processing, some
start time. For example, the solution crosses the 10-centimeter parameters are estimated while others are eliminated via estimates
threshold within 30 minutes on several days, but also takes more derived from a separate process without multiple solutions
time to do so on other days. Shown in FIGURE 2 are the average (in contrast to network RTK); therefore, providing integrity
weekly convergence time series for four different IGS tracking information for PPP single-receiver estimates is that much more
stations (separated by 1,000 kilometers or more and therefore important. A particular industrial service provider has clients
subject to different satellite geometries). Significant differences who are willing to pay for two independent solutions: PPP and
in positioning accuracy and convergence time exist, likely caused long-range RTK (with float ambiguities). The independent
by varying satellite geometry and/or station-specific tracking solutions can be compared to judge their accuracies.
conditions, such as the multipath environment. Obviously, post-fit residuals from a PPP solution can be
Accuracy. The primary factors that limit the accuracy of analyzed to detect individual measurement outliers or more
PPP are the limited precision of current precise orbit and clock significant problems. Such assessment of residuals should be
products and the effects of unmodeled error sources. PPP is standard practice. More complex examples of integrity monitoring
able to provide few-centimeter-level results in static mode and exist in other GNSS applications and should be considered for
decimeter-level results in kinematic mode. Shown in FIGURE 3 are PPP processing. Potentially, a type of receiver autonomous
the positioning results of a high-quality, 24-hour, static dataset integrity monitoring (RAIM) could be implemented to help
from an IGS station (Algonquin Park, or ALGO) using the Jet screen PPP estimates. This would be a straightforward design,
Propulsion Laboratory (JPL) real-time orbit and clock products. and would provide users with additional confidence in their PPP
It can be seen that the coordinate estimates could converge to solutions beyond covariance estimates and post-fit residuals.
the centimeter level within 30 minutes. After convergence, More elaborate, it may be possible to contemplate a Wide Area
all position coordinate components are accurate at the sub- Augmentation System-type of state-space grid error approach to
centimeter level. The positioning-accuracy statistics (root mean evaluate PPP orbit and clock correction products, which could be
square, bias, and standard deviation) are given in TABLE 1. Shown specified when generated from tracking stations and integrated
into the PPP processing.
Latitude Longitude Height
RMS 0.9 1.0 0.7
Potential Improvement
Bias 0.8 0.3 0.0
Having looked at some of PPP’s limitations, let’s see how this
STD 0.3 0.9 0.7
technique might be improved.
▲ TABLE 1 Accuracy statistics for static positioning results (see
Figure 3) after filter convergence (in centimeters) Ambiguity Resolution. Could convergence time be improved
Latitude Longitude Height through stochastic model refinement and use of a higher sampling
RMS 2.8 6.8 4.9 rate? No. It is only when we can exploit the integer property of
Bias ⳮ0.2 ⳮ1.5 ⳮ1.5
ambiguities in PPP that we will have the potential to reduce
STD 2.8 6.7 4.6 convergence time to several minutes or even several seconds. In
▲ TABLE 2 Accuracy statistics for kinematic positioning results double-differenced GPS processing, where integer ambiguity
(see Figure 4) after filter convergence (in centimeters) resolution has been widely utilized, the double-differenced
46 GPS World | April 2009 www.gpsworld.com
 Algorithms & Methods | INNOVATION

Latitude
Latitude Longitude
Longitude Height
Height
3.0
2.4
ambiguity parameters are integers, which can render the position
1.8

Position error (meters)

solution accuracy to the few-centimeter level or better after they 1.2
have been fixed to their correct integer values. For PPP with 0.6
undifferenced observations, the ambiguity parameters are not 0.0
integers as they are corrupted by the initial fractional phase bias -0.6

in the GNSS satellites and receivers. -1.2

-1.8
Significant progress is being made in understanding the char-
-2.4
acteristics and the estimation of the abovementioned initial phase -3.0
biases. Recent undifferenced ambiguity resolution approaches 565200 567600 570000 582400 574800 577200 579600
GPS Time (seconds)
involve a new model or reformulation of the ionosphere-free code
▲ FIGURE 4 Positioning using an aircraft dataset
and carrier-phase observation equations, which, when combined
with the widelane-phase/narrowlane-code observable, permits reported that 50 percent of the horizontal positions were at the
resolution of the 86-centimeter and 11-centimeter ambigui- 2-centimeter level after 10 minutes using ambiguity resolution.
ties. The nature of the new model implies a clock, or clock-like, Integration with RTK. Integration of PPP with network
parameter estimated for each observable for both satellites and RTK techniques may lead to improved position accuracy and
receivers. When applied to static PPP, there is little improve- performance, particularly a reduction in convergence time. Some
ment provided by ambiguity resolution at the end of a 24-hour researchers have already started to investigate such an integration,
period. What ambiguity resolution does provide is the ability to and one commercial vendor has developed a global differential
reach similar levels of accuracy within much shorter observation positioning system. As PPP can be an efficient alternative to
periods. With 30-second test data, after 60 minutes 90 percent RTK in certain applications, it is expected that more work will
of horizontal positions are at the two-centimeter level or better be carried out to investigate the seamless integration of PPP and
with ambiguity resolution, compared to 10 centimeters with- RTK methods.
out ambiguity resolution. In test data sets processed, researchers Integration with INS. The integration of stand-alone

Ge
Gett info rme d.
informed.
Network
Ne with industry
two rk with indus try leaders.
le ade rs .
Learn
Le what yo
arn what your
ur competition
c o mpe titio n isis doing.
do ing .
Attend
Atte the indus
nd the industry’s most
try’s mo s t respected
re s pe c te d conference.
c o nfe re nc e .

Details @ www.gps-wireless.com
www.gpsworld.com April 2009 | GPS World 47
INNOVATION | Algorithms & Methods
Product Accuracy Interval Latency
Final Orbits <5 cm 15 min ~13 days
Clocks <0.1 ns 5 min
Predicted Orbits ~10 cm 15 min Real time
Clocks ~5 ns 15 min
L TABLE 3 IGS orbit and clock products.
and double-differenced GNSS and INS has been extensively
investigated in the past as the coupling has benefits such as
improved cycle-slip detection, smoothed receiver trajectory, and
increased reliability. An integrated PPP GPS/INS system has been
developed to support georeferencing in airborne mapping, which
offers similar performance to a differential GPS/INS system.
An integration of PPP with INS can also reduce re-initialization
time, since INS can supply accurate position and velocity
information during short periods and subsequently reduce the
position convergence time. This is particularly important for
real-time kinematic applications, as frequent signal blockages
are common in the field.
Ingestion of Precision Atmospheric Models. A tropospheric
parameter unknown is usually estimated along with the
position and ambiguity parameters from the measurements in
dual-frequency-receiver-based PPP, while several ionosphere
parameters are estimated in single-frequency-receiver-based PPP.
Ingestion of precision atmospheric models can reduce the total
number of unknown parameters that need to be estimated from
the measurement model, potentially obviating the need for noise-
propagating linear combinations of observables, and potentially
improving positioning performance. As a result, the approach
could reduce the convergence time of PPP. The challenge is that
range corrections from such measurement and physics-based
atmospheric models have to be very accurate, with accuracies
better than a few centimeters.
Infrastructure Issues
Infrastructure refers to the satellite orbit and clock information
products (and potentially, in the future, satellite bias terms for
ambiguity resolution) being generated and used in PPP parameter
elimination schemes, and the information and processes related to
the collection, generation, and dissemination of these products.
Availability of Precise Orbits and Clock Offsets. Precise
orbit and clock products have improved significantly in recent
years. Although advanced modeling and sophisticated software
are required for their generation, post-processed and predicted
products are freely available over the Internet. As such, PPP pro-
cessing can be performed by anyone who can develop the process-
ing software or access one of several online processing engines.
Precise Orbit and Clock Prediction Accuracy. As can be
seen in TABLE 3, post-processed (final) orbit and clock IGS
products are produced at quite a high level of accuracy. However,
improvements in IGS predicted products for real-time usage is
desirable to marginally improve PPP solutions. The greatest
disparity between the final and predicted products is the 50 times
worsening of the clock product coupled with a three times
enlarging of the data rate. Comparable positioning accuracy,
48 GPS World | April 2009
however, has been demonstrated using IGS final products and
JPL real-time orbit and clock correction products.
Real-Time Dissemination of Predicted Orbits and Clocks.
The production of orbit and clock information for real-time
processing, hence prediction of quantities, is a major focus of
current research efforts. The IGS, for example, has been studying
the generation and dissemination of real-time data products
for the past few years and has recently begun a real-time data
products pilot project. A number of other institutes are now
providing predicted orbits and clocks for, amongst other uses,
real-time PPP.
Dissemination of products in the form of corrections can be
done in a variety of ways that can primarily be grouped into
satellite-based or Internet-based. Satellite-based transmission
is the usual choice of commercial service providers, as the
correction signals can be received by an antenna built into the
GPS receiver package. Internet-based correction transmission
can be made with much lower cost and therefore has been used
as the model for academic prototypes, with great potential to
be widely adopted in applications. A common dissemination
protocol is the Networked Transport of RTCM via Internet
Protocol (NTRIP).
The manner of development of PPP and PPP infrastructure
leads to an interesting question of provider model: free versus
paid real-time corrections. That is, an academic/government
model of infrastructure support, or a commercial model of paid
service provider.
Reference Frame. Satellite orbit and clock products refer to
a particular realization of the International Terrestrial Reference
Frame. This realization depends on data from numerous GNSS,
very long baseline interferometry, and satellite laser ranging
stations distributed heterogeneously around the world, and the
non-uniform weighting of the varying-length data records from
each station. When a new version of ITRF is established and
published, and data products redefined with respect to that frame,
PPP user coordinates are slightly changed. The end-user must
be aware of this metadata and, furthermore, that the coordinates
generated by PPP are referred to ITRF and that a coordinate
transformation is necessary to bring the PPP solution into the
“flavor” of coordinates the user requires such as the Canadian
Spatial Reference System implementation of the North American
Datum 1983 (NAD83[CSRS]).
Effects of GNSS Evolution. GPS-only PPP has its limitations,
such as an insufficient number of visible satellites due to signal
blockages and insufficient reliability for safety-of-life applications.
An integration of GPS with other navigation systems, such as the
Russian GLONASS or the future European Galileo and Chinese
COMPASS systems, could provide many more observations and
is expected to have a significant impact on position accuracy,
reliability, and convergence time of PPP. Several researchers
have compared GPS-only and GPS/Galileo-based PPP using
simulations. The results demonstrate that a combined system
can reduce convergence time by half over GPS-only PPP. For
www.gpsworld.com
 MARKETPLACE

To place an ad in this section, contact Candi Eager, 913-254-9635, E-mail: ceager@questex.com

DRAPER LABORATORY

𰀦𰆌𰅽𰅵𰀃𰄐𰅽𰅶𰄐𰄞𰆉𰆚𰀃𰆚𰅽𰀃𰆉𰆌𰅽𰆚𰅽𰆚𰇇𰆉𰄞𰍕𰀃𰇁𰄞𰀃𰄞𰅶𰅐𰅝𰅶𰄞𰄞𰆌𰀃𰆐𰅽𰅯𰆵𰆟𰅽𰅶𰆐𰀃𰆚𰅽𰀃𰆉𰆌𰅽𰄏𰅯𰄞𰅵𰆐𰀃𰅝𰅶𰀃𰁞𰆉𰄂𰄐𰄞𰍕𰀃
𰀘𰄞𰄨𰄞𰅶𰆐𰄞𰍕𰀃𰀯𰅶𰆚𰄞𰅯𰅯𰅝𰅐𰄞𰅶𰄐𰄞𰍕𰀃𰀧𰄞𰅽𰆐𰆉𰄂𰆟𰄂𰅯𰍕𰀃𰀑𰅝𰅽𰅵𰄞𰄚𰅝𰄐𰄂𰅯𰍕𰀃𰄂𰅶𰄚𰀃𰀜𰅶𰄞𰆌𰅐𰇇𰀃𰁞𰇇𰆐𰆚𰄞𰅵𰆐𰍘
𰀒𰅚𰄂𰅯𰅯𰄞𰅶𰅐𰅝𰅶𰅐𰀃𰅽𰆉𰆉𰅽𰆌𰆚𰆵𰅶𰅝𰆟𰄞𰆐𰀃𰄨𰅽𰆌𰀃𰆋𰆵𰄂𰅯𰅝𰄮𰄞𰄚𰀃𰄐𰄂𰅶𰄚𰅝𰄚𰄂𰆚𰄞𰆐𰍗
𰀃 𰍻𰀃𰀧𰁗𰁞𰀃𰄂𰅶𰄚𰀃𰁅𰄂𰇀𰅝𰅐𰄂𰆟𰅽𰅶𰀃𰁤𰄂𰆐𰅬𰀃𰀾𰄞𰄂𰄚𰄞𰆌𰆐
𰀃 𰍻𰀃𰀧𰁗𰁞𰀃𰄂𰅶𰄚𰀃𰁅𰄂𰇀𰅝𰅐𰄂𰆟𰅽𰅶𰀃𰀜𰅶𰅐𰅝𰅶𰄞𰄞𰆌𰆐
𰀃 𰍻𰀃𰁚𰀦𰍬𰀒𰅽m𰅵𰆵𰅶𰅝𰄐𰄂𰆟𰅽𰅶𰆐𰀃𰀘𰄞𰆐𰅝𰅐𰅶𰄞𰆌𰆐𰀃𰄂𰅶𰄚𰀃𰀾𰄞𰄂𰄚𰀃𰀜𰅶𰅐𰅝𰅶𰄞𰄞𰆌𰆐
𰁙𰆵𰄂𰅯𰅝𰄮𰄞𰄚𰀃𰄂𰆉𰆉𰅯𰅝𰄐𰄂𰅶𰆚𰆐𰀃𰅵𰆵𰆐𰆚𰀃𰄏𰄞𰀃𰁨𰍘𰁞𰍘𰀃𰀒𰅝𰆟𰇌𰄞𰅶𰆐𰀃𰅽𰆌𰀃𰆉𰄞𰆌𰅵𰄂𰅶𰄞𰅶𰆚𰀃𰆌𰄞𰆐𰅝𰄚𰄞𰅶𰆚𰆐𰀃𰇁𰅝𰆚𰅚𰀃𰄂𰄏𰅝𰅯𰅝𰆚𰇇𰀃𰆚𰅽𰀃
𰆐𰄞𰄐𰆵𰆌𰄞𰀃𰀘𰅽𰀘𰀃𰄐𰅯𰄞𰄂𰆌𰄂𰅶𰄐𰄞𰍘𰀃𰀘𰆌𰄂𰆉𰄞𰆌𰀃𰅽𰄫𰄞𰆌𰆐𰀃𰄂𰅶𰀃𰄞𰇆𰄐𰄞𰅯𰅯𰄞𰅶𰆚𰀃𰄐𰅽𰅵𰆉𰄞𰅶𰆐𰄂𰆟𰅽𰅶𰀃𰄂𰅶𰄚𰀃𰄏𰄞𰅶𰄞𰄮𰆚𰆐𰀃𰆉𰄂𰄐𰅬𰄂𰅐𰄞𰍘
𰀄𰆉𰆉𰅯𰇇𰀃𰅽𰅶𰅯𰅝𰅶𰄞𰀃𰄂𰆚𰍗𰀃𰇁𰇁𰇁𰍘𰄚𰆌𰄂𰆉𰄞𰆌𰍘𰄐𰅽𰅵𰀃
𰀜𰀜𰍬𰀄𰀄𰀃𰄞𰅵𰆉𰅯𰅽𰇇𰄞𰆌 #*
Engineering Innovations

example, we demonstrated that combined
GPS/GLONASS PPP greatly improves
positioning accuracy and reduces
convergence time (see FIGURES 5 and 6).
These improvements are dependent on
the enhanced level of satellite availability
and geometry for position determination.
Issues such as interoperability and
compatibility, however, must be addressed
for successful integration of data from
hybrid navigation systems.
PPP versus RTK
Given the previous discussions about
the capabilities and potential of PPP, an
obvious topic to conclude this article is a
comparison of PPP and the industry-stan-
dard RTK technique, and to answer the
question: Can PPP ever replace RTK? To
address this overall issue, we pose a series
of logically ordered questions. Some que-
ries may be answered quite readily, while
others can only be partially addressed.
www.gpsworld.com
1. Can PPP algorithms, data and op-
eration be improved to the point where
the technique obtains the same level of
Compact antennas for handheld &
performance as RTK? And, if so, what Mobile GPS/SatCom
specific improvements are required for
Quadrifilar
PPP to perform like RTK? Helical Antennas
This result can be seen as a potential *10mm x 20mm
*Gain -1.3 dBic
final objective of PPP algorithm research *0.2 dB Axial Ratio
— and a recasting of the overall question. Microstrip
All of the measurement strength of the Patch Antennas
*Available in 5 sizes
undifferenced PPP observables is used *Gain 6.3 dBicv
to determine: 1) few-centimeter-level Now hiring RF & antenna engineers. BSEE
and US citizen prefered. Send resumes to
position estimates, 2) with a few seconds jobs@maxtena.com
worth of measurements, 3) without the
1-877-MAXTENA
need for a reference station. The utility 1715 Pratt Drive, Suite 2800
Blacksburg, VA 24060, USA
of such a solution would be significant www.MAXTENA.com
(see below).
The third characteristic (of removing
the need for reference stations) is the one research results related to ambiguity
that would make PPP so appealing as resolution of undifferenced observables.
compared to RTK. The first characteristic The potential exists to isolate and
(of attaining centimeter-level positioning estimate initial fractional phase biases in
accuracy) may be possible, given recent order to isolate true integer ambiguities,
April 2009 | GPS World 49
INNOVATION | Algorithms & Methods

3
Position error (centimeters)

2 East
North
1 Up
RTK, or Differential GPS?
0
We can speculate that PPP would f irst be used as a
-1
complementary solution to RTK in positioning and navigation
-2
-3
work. But as the approach gained acceptance, it could replace
6:40 7:10 7:40 RTK. The only caveat here is that a significant level of integrity
Time (hours:minutes)
▲ FIGURE 5 Simulated GPS-only PPP solution convergence period would have to be guaranteed along with accurate PPP solutions
and accuracy performance to gain industrial acceptance.
3
5. How would science, industry, and society be affected?
Position error (centimeters)

2 East RTK-like PPP would positively affect activities in all of these

North
1 Up areas. Relieved of RTK-baseline constraints, users would be able
0 to perform few-centimeter-level positioning almost anywhere,
-1 though performance would be similar in urban areas where
-2
RTK networks are already established. That said, a two-system
-3
6:40 7:10 7:40 solution would enable new applications.
Time (hours:minutes)
▲ FIGURE 6 Simulated GPS/GLONASS PPP solution convergence
period and accuracy performance Conclusion
PPP is a work in progress. Already providing significant benefits
without over-parameterizing the processing model. The second to many user communities, PPP will continue to evolve and
characteristic (of attaining the desired accuracy with a few seconds improve as researchers develop innovative techniques to finally
of data) is perhaps the most challenging. Given the inherent solve the ambiguity resolution problem, and then proceed to
weaker measurement models of PPP versus RTK (that is, less further improve accuracy and convergence period — further
data in PPP), it will be difficult to estimate the correct biases and expanding the technique’s utility.
ambiguities, aside from the problem of performing this resolution
quickly. Meeting the initialization and re-initialization challenge Acknowledgments
will be the most difficult hurdle for the PPP technique to receive The authors acknowledge support from the Geomatics for In-
greater industrial acceptance for real-time applications. formed Decisions Network of Centres of Excellence and the
2. Can this objective be reached in the near future? In a cost- Natural Sciences and Engineering Research Council of Canada.
effective way? In a practical manner? This article is based on the paper “Current Status of Precise Point
If the objective of parity between techniques can be reached, Positioning and Future Prospects and Limitations” presented at
it will require significant further algorithm development and the 24th General Assembly of the International Union of Geod-
perhaps more observables and independent data to reduce the esy and Geophysics held in Perugia, Italy, July 2–13, 2007.
processing-filter convergence period. This latter requirement
SUNIL BISNATH is an assistant professor in the Department of Earth and
would delay implementation of an RTK-like PPP processor. Space Science and Engineering at York University in Toronto, Canada.
No infrastructure changes are required to estimate receiver He has a Ph.D. degree from the University of New Brunswick, and his
and satellite code and phase biases, and the added complexity research interests include precise GNSS positioning and navigation
algorithms and applications.
should not increase costs to service providers. If, as some research
suggests, few-minute convergence periods are possible, we are YANG GAO is a professor in the Department of Geomatics Engineering
hesitant to state that this limitation will be tolerated by many at the University of Calgary (U. of C.). He has a Ph.D. degree from
industries other than the ones currently using PPP. U. of C. and his research expertise includes both theoretical aspects
and practical applications of satellite positioning and navigation sys-
3. Do we even want PPP to work like RTK? tems. His current research focuses on high-precision GPS positioning,
Though it may seem like an odd question, given the context differential GPS networks, and mobile information management.
of the discussion, it is a reasonable question to ask if, at least
theoretically, PPP can perform at the level of RTK. The two MORE ONLINE
approaches have been developed independently, for different
purposes. This fact makes them very useful as independent, Further Reading
mutual integrity checkers for some scenarios. For references related to this article, go to gpsworld.com and click
on Innovation under Resources in the left-hand navigation bar.
4. If PPP can work like RTK, would it replace RTK, network
COPYRIGHT 2009 QUESTEX MEDIA GROUP, INC. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical including by photocopy, recording, or information storage and retrieval
without permission in writing from the publisher, Questex Media Group, Inc. Authorization to photocopy items for internal/educational or personal use, or the internal/educational or personal use of specific clients is granted by Questex Media Group, Inc. for libraries and other
users registered with the Copyright Clearance Center, 222 Rosewood Dr., Danvers, MA 01923, 978-750-8400 fax 978-750-4470. For uses beyond those listed above, please direct your written request to questexpermissions@reprintbuyer.com or 800-494-9051, Ext. 100.
PRIVACY NOTICE: Questex Media Group provides certain customer contact data (such as customers’ names, addresses, phone numbers and e-mail addresses) to third parties who wish to promote relevant products, services and other opportunities which may be of
interest to you. If you do not want Questex Media Group to make your contact information available to third parties for marketing purposes, simply call toll-free 866-344-1315 or 1-847-763-9594 (Outside the US) between the hours of 8:30 am and 5 pm (CT) and a customer service
representative will assist to remove your name from Questex’s lists.
GPS WORLD (ISSN 1048-5104) is published monthly (12 issues/yr) by Questex Media Group, Inc., 306 W Michigan St, Ste 200, Duluth, MN 55802. SUBSCRIPTION RATES: U.S. and possessions — 1 year (12 issues), $67; 2 years (24 issues), $117. Canada
and Mexico — $87, 1 year; $137, 2 years. All other countries — $127, 1 year; $232, 2 years. International pricing includes air-expedited service. Single copies (prepaid only): $7 in the United States, $9 all other countries. Back issues, if available, are $19 in the U.S. and
possessions, $23 all other countries. Include $6.50 per order plus $2 per additional copy for U.S. postage and handling. Periodicals postage paid at Duluth, MN 55806, and additional mailing offices. POSTMASTER: Please send address changes to GPS WORLD,
P.O. Box 1270, Skokie, IL 60076-8270, USA. Canadian G.S.T. number: 840 033 278 RT0001, Publications Mail Agreement Number 40017597. Printed in the U.S.A.

50 GPS World | April 2009 www.gpsworld.com

With a tradition of innovation, NavCom continues to leave its footprint on
the evolving GNSS industry. Propelled by our award-winning engineering
team, our industry leading technology includes the StarFire™ network – the
world’s first global satellite-based augmentation system (GSBAS) – RTK
Extend™ and Ultra RTK™. This unique expertise along with NavCom’s precise
positioning products allows our partners to create innovative GNSS solutions.
As John Deere’s technology steward, NavCom continues to introduce
groundbreaking solutions, enabling our business partners to take on larger
projects, improve their efficiency – and make their own history.
NavCom – Here today, here tomorrow.
To learn more, call us at +1-310-381-2000.

w w w . nav com t ech. com

GPS
WOULD BE LOST
WITHOUT US.
IMPROVEYOURGPS.COM