Free eBook: The Risks and Limitations of GNSS Live-Sky Testing
Live-sky testing is a quick way to assess the basic operation of GNSS-enabled devices—but it’s also a severely limited one, and a practice that could put both the quality of your products and the reputation of your brand at risk.
Download a free eBook to discover:
- The inherent lack of control, repeatability and scalability of live-sky testing
- The potential impact on time to market, and on development and support costs
- How GNSS simulation resolves these problems, helping your products perform as intended
Simply enter a few details opposite to receive your free eBook—and happy reading!
About Spirent
Spirent has been the global leader in GNSS testing for near 30 years. Spirent delivers navigation and positioning test equipment and services to governmental agencies, major manufacturers, integrators, test facilities and space agencies worldwide.
The risks and limitations
of GNSS live sky testing
within a production environment
Challenges for the unfamiliar
The addition of location awareness to all manner of high-volume
consumer products brings with it a number of challenges for
manufacturers that are not necessarily familiar with satellite
navigation technology.
And these challenges do not end once the product is designed and
debugged.
When the product moves into production, the challenge becomes
how to ensure that every single unit leaving the factory will perform
exactly as it should. Because any malfunctioning location aware
product will very quickly be found out ¬nullas soon as it gives its users
an inaccurate null on their position, or delivers navigation instructions
that are patently wrong.
Page 2SPIRENT eBook
nullanufacturers of consumer products
from cars to mobile handsets have
become all too aware of the power of
the disgruntled user. nullnline forums
and todaynulls social media provide the
dissatisnulld customer with the means to tell the world that any given
product does not do nullexactly what it says on the tinnull
nulld while such media can offer valuable opportunities to improve
customer interaction, recurring negative comment can do irreparable
damage to any brandnulls reputation for reliability.
The dangers
of failure
Page 3SPIRENT eBook
Quality control
nullearly, the only way to ensure that products released to the market
are null for purpose is through functional testing at the end of the
production line. nulld in this respect, the nullnull receiver is no different
from any other function of the end equipment.
The challenge then becomes how best to integrate functional testing
of the nullnull receiver with the other tests performed on the product.
nulld that means devising a test strategy that will exercise the
equipment sufnulliently to prove its correct nullr incorrectnullfunctionality
in the minimum possible test time.
Page nullSPIRENT eBook
Functional testing
null with any functional testing, the operation will involve applying a
known stimulus and observing the response. nullf the response is within
acceptable bounds the product passes the testnullif the response is
outside those bounds the product fails.
The response to the stimulus may simply be a trigger signal or an
alarm, or it may be some variable output that can be measured for
the purposes of the test. But whatever the expected result, it is
essential that the stimulus is accurate and repeatable so that the
response can be gauged against a known standard.
Page nullSPIRENT eBook
The tests required
The exact tests required will depend entirely on the intended
use of the end equipment. nulld while it will not be necessary to
exercise all the functions of the nullnull receiver, the tests need to
ensure that the receiver will function as required when it is
used nulln angernull
nullo while static navigational accuracy may well be the simplest
test to perform, other aspects such as time to nullst null nullTnullnull
or renullcnullisition time might well be important to the overall
operation of the equipment.
Page nullSPIRENT eBook
Supplying the stimulus
When integrating nullnull tests with other acceptance tests, the nullst
nullnd rather obviousnullproblem is that the test area is not enullctly
a GNSS friendly environment. The roof of the production area will
almost certainly block the ingress of signals from any nullnull system,
and there will inevitably be a great number of interfering null signals
being renullcted around the test area.
The nullst of these problems is relatively easy to solvenullthe nullnull
signals can be nullcapturednullby an outdoor receiver and relayed to the
test area. nullowever, while relaying nullive-skynullsignals might seem to
offer a cost-effective solution for production testing of nullnull-enabled
equipment, it does have many signinullant limitations.
Page nullSPIRENT eBook
The limitations of linull sky testing
nullve sky testing describes a scenario where the null signals being sent
to the nullnullreceiver in the test system are from navigation satellites
overhead at that time. The antenna which is receiving the signals from
space is connected to the nullnullreceiver by cable or Bluetooth and is
typically mounted on the roof of a building or nullted to a test platform
such as a motor vehicle, train, ship or aircraft.
nullimple diagram herenull
Page nullSPIRENT eBook
nullf using a test platform null.g. an aircraftnulllive sky testing allows in-
situ analysis of system performance including elements such as
antenna placement and electronic interference. nulln addition, by
using splitters and attenuators a single nullnullantenna, mounted
for example on the roof of a building, can feed multiple test
rigs. This simplicity of installation and operation, which can
easily be achieved by nullnullnovices, is the main reason why some
organinulltions initially choose this method.
Page nullSPIRENT eBook
nullowever, for those requiring a rigorous nullnulltest environment,
live sky testing has some serious limitations including a lack of
repeatability, control and scalability. nullepending on the application,
it can also turn out to be a resource intensive test method with
potentially negative impacts on time to market, development and
support costs. There is also the issue that in some countries re-
radiation of nullnull signals is illegal.
Page nullSPIRENT eBook null
Repeatanulllity
By denullition, live sky testing is not a repeatable test method
because the nullnullconstellation constantly changes. nullor example,
if using a static antenna the dilution of precision nullnullnullof the
navigation solution will be observed to vary as satellite anullmuths,
elevations and signal strengths change. When using a nullnullreceiver
on a moving test platform additional variables are introduced such
as the slew, pitch and yaw of the nullnullantenna relative to the
satellites along with signal attenuation, renullction and refraction
from overhanging trees, adnullcent buildings etc.
What all this means in practice is that the nullnullconditions causing
an error during one test run cannot be reliably and repeatedly
replicated during subsequent live sky testing runs.
Page nullSPIRENT eBook
nullne technique to overcome this serious limitation to live sky testing
is recording the nullnulldata during the nullst test and replaying the same
nullnulldata in the lab for subsequent tests. This does allow a measure
of repeatability to be introduced to the testing regime but does not
however resolve the issue of control.
Page nullSPIRENT eBook null
Control
nullecording and replaying nullnulldata to a test system allows you to
replicate the symptoms of an error, for example a position nullmp, but
it doesnnull give you an answer regarding the root cause
of the problem. Was it the introduction of a
particular satellite to the solutionnullWas it
caused by multi-path or signal blockage in
a particular directionnullnulloes it happen under
any ionospheric conditions or only the
conditions occurring at the time the data
was collectednull
SPIRENT eBook
nullnly a simulator offers full control over all relevant nullnullvariables,
facilitating a test regime which can isolate the root cause of nullnull
induced errors. null simulator also allows hypothetical scenarios to be
created and can consistently create signal errors and events null.g.
nullnullnullwhich may only occur sporadically or randomly in the real-
world. nullinally, unlike live sky testing or record and replay, simulators
are a fully scalable testing solution.
Page nullSPIRENT eBook null
Scalanulllity
nullve sky testing lacks scalability in three important respects. nullirst,
it does not adequately test products for global marketsnullsecond,
it cannot test for future signals and satellites, andnullthird, it is not
suitable for a test regime or test standards applied across multiple
development sites.
nullve sky testing is based on a single scenario nullhe constellation and
environment at the time of the testnulland a single location. null single
scenario and location can never renullct the diversity of conditions in
which a nullnullsystem or application may have to operate.
Page nullSPIRENT eBook
The alternative is multiple live sky tests based on different
constellations null.g. high nullnull versus low nullnull, environments null.g.
urban, rural, forestnulland locations null.g. high latitudes, equatorialnull
however, as discussed later, this approach has its own set of issues in
terms of time to market and costs often making it impractical.
nullnullis one of the six nulllobal nullavigation nullatellite nullystems nullnullnull
and four nullatellite Based nullugmentation nullystems nullBnullnullwhich will
be available by the middle of this decade. There are already over null
nullnull and nullnull satellites in operation nullncluding null from nullnull and
more than nullnullare planned. nullnull incorporates both the existing nullnull
null and null frequencies along with null, nullnull null, null, nulla, nullb, null and
many more. nullystems tested today using live sky methods therefore
have a real risk of unexpected errors as the full nullnull constellation
comes online.
Page nullSPIRENT eBook null
nullinally, if the requirement is to introduce a common testing regime
across multiple development sites nullhich might be on different
continentsnulllive sky testing does not deliver a standardinulled
method. nullecorded data could be used as a standard but this still
leaves problems with control nulliscussed earliernull time to market,
development and support costs.
Page nullSPIRENT eBook null
Time to marketnulldenulllopment and
support costs
null discussed earlier, the apparent ease of setup and low cost of
live sky testing can initially be attractive to organinulltions seeking to
establish a nullnulltesting regime. nullowever, more often than not, people
going down this route have not calculated the true costs.
nullnsider for example how many scenarios and locations are required
to adequately test a nullnullsystem. nullf we assume that we need ten
scenarios at six locations that would require no less than sixty nullnull
live sky tests - each consuming time, engineering resources and
expense. nullven if the test platform was as simple as a motor vehicle
the total costs of running sixty live sky tests could easily run into tens
of thousands of dollars.
Page nullSPIRENT eBook
Whatnulls more, even if the data is recorded for future replay nullhich
helps nullstify the costnullit has a limited shelf life. nullor as soon as a
new satellite or frequency is launched null.g. null on nullnullnullnullsatellitesnull
historic data technically becomes redundant and new data should be
collected.
nullven with the best of planning, as the development process unfolds
it is likely that additional scenarios andnullr locations will be required.
nullch time this happens it not only increases costs but can add
unanticipated delays as resources have to be shifted and time found
to collect new data.
Page nullSPIRENT eBook null
null similar situation is found during beta testing and after product
release. null customer problem involving a new scenario andnullr location
may require yet more data to be collected to adequately analynulle and
resolve the issue.
null these simple examples demonstrate, depending on the volumes
and geographic spread of markets the nullnullor nullnull product is
destined for, live sky testing can turn out to be a much more
expensive option than people nullst think.
Page 2nullSPIRENT eBook
nullnnull null fooled
nullrtainly, if the only test required is for the equipment to recognise
the presence of nullnull signals, then the relayed signals would be
adequate for this purpose. nullimilarly, if the obnullct of the test is for the
equipment to determine its location from the signals supplied, then
the relayed live-sky signals will sufnulle.
Page 2nullSPIRENT eBook
nullowever, the whole ethos of nullality control is to determine whether
the enullipment produces the enullected response to a known stimulus.
nullrtainly, the location of the test rig will not change, and so the
position calculated by the nullnull receiver should always be the same.
But the nature of the “live” signals that are being relayed will be
changing constantly. The relative positions of the satellites within the
constellation will change; the atmospheric conditions will never be
constant; and other external influences such as interference will not
be consistent. So much for applying a “known” stimulusnull
Page 22SPIRENT eBook
The initial attraction of live sky testing is the ease and cost of
setup. nullowever, as we have demonstrated above, this method has
serious limitations in terms of repeatability, control, scalability, time
to market, product development costs and product support costs.
nullthough the record and replay method overcomes some of these
limitations it cannot provide the rigorous test regime that a nullnull
simulator can offer.
Page 23SPIRENT eBook
The relianulle alternatinull
null nullnull simulator provides the ideal alternative to any form of nullivenull
signals for testing any nullnull receiver. nulld while the simulator that
would typically be used in characterising the design would need to
be a fully-featured multichannel instrument capable of simulating the
signals of an entire constellation of satellites, the simulator used for
production testing can be a relatively simple singlenullhannel unit.
The important aspect here is that the simulator will produce enullctly
the same stimulus each and every time. Therefore any variation in the
response of the equipment under test can easily be observed. nulld if
the response falls outside acceptable bounds then the equipment will
be renullcted.
Page 2nullSPIRENT eBook
Bringing it all together
nullf course, it is highly unlikely that the nullnull receiver will be the
only part of the equipment that requires acceptance testing. nullo it is
important to choose a nullnull simulator that can be integrated with
other test equipment in order to create a complete functional test
system for the whole equipment.
The combined test setup will speed product throughputnull minimising
test time and enabling complex nullf-thennulltest scenarios that can
exercise multiple functions in a single software-controlled test routine.
Page 2nullSPIRENT eBook
Reputation presernulld
By integrating simulator-based nullnull receiver testing with other
functional testing as part of a complete quality control procedure,
manufacturers can be connullent that their locationnullnabled products
will perform as designednull helping maintain consumer connullence in
their hard earned brand reputations.
Page 2nullSPIRENT eBook
We hope you found this Spirent Live Sky Testing E-Book
of interest.
We are continually adding new content to our website
on a regular basis. Bookmark this link:
www.spirent.com/Positioning-and-Navigation/Pnullnullideonullibrary.aspnull
nullisit the Spirent GNSS Blog, there are currently
more than null posts with nullto nullnew posts added
per week. nulltch up on whatnulls new:
www.spirent.com/Blog/Positioning
Need more informationnull
gnss-solutionsnullpirent.com
nullharenull
nullacebook nullnkednulln nullwitter nullechnorati nulloogle Bunull nulligg nullelicious nulleddit nulltumbleupon
nullnullnullnull
Contact null
Spirent
nullnullnullnull nullnullnull
globalsalesnullpirent.com
www.spirent.comnullositioning
Spirent nullederal Systems
null nullnullnullnullnullnull
infonullpirentfederal.com
www.spirentfederal.com
Got a smartphonenull
Scan the null
nullde for more
information
