Journal of Global Positioning Systems (2004)
Vol. 3, No. 1-2: 154-166
Limitations of Pseudolite Systems Using Off-The-Shelf GPS Receivers
Mustafa Ozgur Kanli
School of Surveying and Spatial Information Systems, The University of New South Wales, Sydney, Australia
e-mail: [e-mail] Tel: + 61-2-6229-1780; Fax: + 61-2-6229-1778
Received: 15 Nov 2004 / Accepted: 3 Feb 2005
Abstract. Pseudolites (PLs) are ground-based
transmitters that transmit GPS-like signals. They have
been used to test GPS system elements and to enhance
GPS in certain applications by providing better accuracy,
integrity and availability through the use of PL signals in
addition to the GPS signals. PLs are also a promising
technology for providing positioning in indoor, high
multipath environments where GPS signals are generally
unavailable or severely attenuated and of questionable
quality. In experiments to date, researchers have almost
exclusively used PLs that transmit C/A code on L1/L2 in
order to use existing off-the-shelf GPS receivers. This is
because no hardware modifications to the GPS receiver
are necessary and only minor changes to the receiver
firmware are needed to track a PL’s signal. However,
there are some fundamental issues that limit the
effectiveness of a PL system using C/A code on L1/L2.
These include the legality of transmitting on L1/L2,
cross-correlation between PL and GPS signals, saturation
of GPS receiver front-ends, and the limited multipath
mitigation offered by C/A codes. When combined with
other problems inherent to all PL systems such as near-
far, multipath, and synchronization, the issues in using
L1/L2 C/A code PL systems further complicates the
design and deployment of such systems and places limits
on its operational effectiveness. This paper presents the
issues which limit PL systems that use GPS hardware and
explores the impact of these issues on some common PL
applications.
Key words: pseudolite, cross-correlation, jamming.
1 Introduction
The concept of using ground-based GPS signal
transmitters originated from the time when GPS was first
being developed. These transmitters, which came to be
known as Pseudolites (PLs), enabled developers to test
GPS concepts before the first satellites were even
launched. Since then, PLs have been found to improve
geometry for accurate positioning, especially in the
vertical component. They provide additional ranging
signals to augment existing GPS signals, increasing
availability. In some environments, such as indoors, GPS
signals are heavily attenuated and are of questionable
quality. In these cases, PLs have been used to provide
alternative ranging signals. A comprehensive summary of
PL technology, including PL designs and applications, is
provided by Wang (2002). There, the reader will also find
many helpful references.
Almost all purpose-built PLs transmit on the L1
frequency. Some also transmit on L2, or even both. Those
that are simply comprised of a signal generator may
transmit on any range of frequencies. Most L1/L2 PLs
typically use spreading codes from the same family as the
C/A codes used in GPS satellites. This practice enables
the PL to closely resemble a GPS satellite, and this is
important to allow existing off-the-shelf (OTS) GPS
receivers to track a PL’s signal. Only slight changes to a
GPS receiver’s firmware is required to achieve this. First
is a modification to the receiver’s search database,
allowing it to look for local PLs as well as GPS satellites.
If the PL modulates data onto its signal, software to
correctly interpret this data also has to be written. Some
other thresholds and assumptions may also need to be
adjusted. Importantly, no hardware changes to a
receiver’s correlators or front-end circuitry are needed.
This practicality simplifies experimentation and leaves
more time to be spent on research.
Unfortunately, this convenience comes with a price. The
L1 and L2 frequency bands are protected and reserved for
radionavigation; legal issues hinder unlicensed
transmissions in these bands. PL signals are often
considerably stronger than GPS signals. This can result in
jamming, which will deny GPS to other nearby receivers
that don’t participate in PL navigation. Typical OTS GPS
receivers may also be overwhelmed by the strength of PL

Kanli: Limitations of Pseudolite Systems Using Off-The-Shelf GPS Receivers 155
signals; they expect weak GPS signals. Saturation of the
front-end can result, which will decrease accuracy,
reliability and introduce other complications. For indoor
PL navigation, the C/A code chipping rate used in GPS
does not sufficiently limit range errors due to multipath.
These factors contribute to other complications that can
affect the operational effectiveness of PLs.
This paper will begin with a discussion of issues inherent
to all PL systems; these are near-far, multipath and
synchronization. Then, issues specific to L1 C/A code
PLs will be presented. A brief look at the impacts of these
issues on some common applications will follow. The
paper will then conclude with a section discussing the
recent developments in PL technology. Other competing
positioning systems will also be briefly discussed, as well
as some possible future developments.
2 Issues in pseudolite systems
2.1 General Issues
This section will describe the general issues confronted
by PL systems.
2.1.1 Near-far
GPS satellites are located in near-circular orbits around
the Earth with a radius of about 26,560 km. This puts the
typical distance between a user receiver and a visible
satellite overheard at about 20,000 km. Changes in this
distance due to typical user and satellite motion are
insignificant compared to the overall separation distance.
Also, the radiation patterns of GPS satellite antennas are
shaped to spread the RF signal almost uniformly over the
surface of the Earth. These factors ensure that a user
receiver can expect to see GPS signal strengths around -
160 dBW from all visible satellites. In contrast, PLs are
located much closer to user receivers so that user
movements can cause significant variations in the
distance between PLs and user receivers. Because of this,
user receivers will see large variations in PL signal
strengths. For example, for an isotropic PL antenna the
signal strength 30 m away will be 24 dB stronger than the
signal strength 500 m away. This is known as the near-far
effect.
The repercussions of the near-far effect on GPS signals
are primarily jamming and limited operating distance. For
instance in a standalone PL network each PL will, within
an area of close proximity, jam the signals from other
distant PLs. Receivers in such a network will observe
varying PL signal strengths at different locations. Off-
the-shelf GPS receivers expect signals from all sources to
be of approximately equal strength. This will only occur
at points equi-distant to all PLs. PL signals can be more
than 30 dB stronger than GPS signals; which is greater
than the worst case code separation (21.6 dB) between
C/A codes. In an integrated GPS/PL system, PLs can thus
potentially jam GPS signals at close proximity. The
output power of the PL can be calibrated to minimize
jamming, but this limits the operating range of the PL.
The most popular solution used to mitigate the near-far
problem, as presented by Klein and Parkinson (1984), is
to pulse the PL signals at fixed cycle rates (typically
10%). This provides a 10 dB improvement in the signal-
to-interference level. Transmitting the signals at a
frequency offset from GPS L1 (while remaining within
the same frequency band as GPS) and using longer
sequence spreading codes have also been suggested as
potential solutions. However, not all off-the-shelf GPS
receivers can accommodate the use of different spreading
codes and/or the frequency offsets. A different approach,
proposed by Madhani et al (2001), uses successive
interference cancellation, where the receiver accurately
regenerates the interfering PL signal before subtracting it
from the total received signal to yield a partially cleaned
version of the received signal. This approach relies on the
ability of the receiver to accurately track both the
interfering signal and the desired weaker signal. In order
to do this, the receiver requires an ADC with sufficient
resolution and an RF front-end with sufficient dynamic
range to accommodate both sets of signals. Most off-the-
shelf GPS receivers, however, don’t have this capability.
Shaping the radiation pattern of PL transmit antennas is
another approach used in the mitigation of the near-far
problem (Söderholm et al, 2001). This approach aims to
provide a near-uniform spread of PL signal strength
covering only a specific desired area, achieved by
appropriate orientation of the transmit antennas of the
PLs. While effective within the desired area, near-far
conditions are still apparent on the fringes of, and outside,
this operational area. Custom antenna designs are
required to provide sufficient coverage of each
environment. Thus, this approach will not be suitable for
dynamic environments such as in shipping yards, where
moving freight containers cause considerable changes to
the landscape. Furthermore, the deployment of PLs is
restricted by the very nature of this near-far mitigation
technique.
2.1.2 Multipath
A pseudo-range measurement involves determining the
propagation time of a ranging signal along a direct line-
of-sight path from its source antenna to a receiver’s
antenna. Due to reflective objects in the environment, the
signals at a receiver’s antenna can be composed of both
direct ranging signals and any number of reflected
(multipath) ranging signals. These multipath signals are

156 Journal of Global Positioning Systems
delayed relative to the direct signal and have differing
amplitude, phase, and polarization, characterized by the
reflecting surface and the number of reflections. Pseudo-
range errors of tens of metres often result from the
presence of multipath signals, which in most cases is the
largest error source. Multipath signals give rise to larger
pseudo-range measurement errors than carrier phase
measurement errors, which are of the order centimetres.
Multipath signals may also destructively interfere with
direct signals resulting in multipath fades. A good
introduction to multipath effects is given by Braasch
(1996).
Unlike near-far, the problem of multipath is an issue for
GPS users as well as for designers of PL systems. The
severity of multipath varies with the environment in
which the positioning signals are applied. As an example,
when PLs are used for augmenting GPS in aircraft
approach and landing, a typical source of multipath is
long delay ground bounce. In this case, multipath signals
are typically delayed by more than a ranging code chip
period. Theoretical texts may show that the multipath
signal will have no effect. However, C/A code correlation
side-lobes provide a way for relatively strong long delay
multipath signals to induce errors of several metres. It
should be noted that a GPS receiver may choose to ignore
satellites at low elevations to exclude the possibility of
ground bounce signals introducing pseudo-range errors.
For PL systems however, all signal sources are almost
always at low elevations significantly increasing the
amount of multipath.
A different case is a PL system for indoor positioning,
where there is a wide range of reflectors in close
proximity of the receiving antenna. Examples include
reinforced concrete structures, home/office furnishings
and antenna mountings (man or machine). These
reflectors generate a prevalence of short delay multipath,
where ranging signals are delayed by less than a ranging
code chip period. Short delay multipath typically has a
greater impact on pseudo-ranges than long delay
multipath, and it is also harder to mitigate its effects. For
example, a single multipath signal with ¼ the amplitude
of the direct signal can induce a 40 m range error (Misra
and Enge, 2001).
Multipath mitigation can be performed on several levels.
First, multipath signals can be attenuated selectively at
the antenna. A choke-ring antenna is one device that can
achieve this (Kunysz, 2003). It consists of concentric
rings of grounded metal around the antenna element.
These attenuate signals that arrive from low elevations
relative to the axis of the metal rings. Choke-ring
antennas are effective in attenuating ground bounce
multipath signals. Other directional antenna designs have
been used to attenuate signals from non-line-of-sight
paths (Stolk and Brown, 2003). Of course, such
techniques are only useful if the multipath signals arrive
from the directions they are anticipated from.
Modelling multipath errors in software in order to
minimize its effects is another type of mitigation. One
approach includes the simulation of different types of
reflectors at a variety of distances from receiver antennas.
Receiver tracking algorithms can then be modelled to
determine the response of the tracking loop to the
reflected signals (Weiss et al, 2003). Modelling multipath
or multipath estimation may be effective for controlled,
static environments but it is unsuitable for dynamic,
indoor and high multipath environments. In such cases,
receivers moving through a reflective environment will
encounter multipath effects that deviate from the model.
Multipath mitigation is also performed at the level of the
receiver’s correlators. Ordinary receivers perform
correlation with a pair of early and late tracking arms
separated by a chip period. These tracking arms straddle
the correlation peak by maintaining the level difference
between the pair at zero. In the presence of a multipath
signal, however, keeping the level difference at zero may
not imply that the tracking arms straddle the true
correlation peak. Decreasing the separation between the
tracking arms is one way to achieve a better result in the
presence of medium delay multipath signals. The Narrow
CorrelatorTM operates in this fashion (Van Dierendonck et
al, 1992). Other correlation techniques such as the
Double Delta correlator, the Early/Late Slope technique
and the Early1/Early2 tracker have also been devised.
These can dramatically decrease the maximum pseudo-
range error due to multipath. Irsigler and Eissfeller (2003)
present a good comparison of the multipath mitigation of
different correlation techniques mentioned here. It is
important to note that techniques like the Narrow
CorrelatorTM require a receiver with a front-end having a
wide bandwidth relative to the chipping rate of the
ranging code used. For C/A code, typical narrow
correlator receivers have pre-correlator bandwidths of 16
MHz. The consequence of having such a wide bandwidth
is that signals from adjacent radio bands are more likely
to cause interference. Also, not all OTS GPS receivers
employ such advanced multipath mitigation techniques.
2.1.3 Pseudolite synchronization
Fundamental to the operation of GPS is satellite clock
predictability. Without precisely synchronized satellite
clocks, precise time transfer and accurate stand-alone
navigation would be impossible. The cesium and
rubidium atomic clocks onboard GPS satellites have
stabilities of the order of 1 part in 1013. To keep the
clocks synchronized, the GPS Operational Control
Segment (OCS) uploads satellite clock corrections to the
satellites at least once a day. These corrections, which are

Kanli: Limitations of Pseudolite Systems Using Off-The-Shelf GPS Receivers 157
part of the navigation message, are used by receivers to
correct for the satellite clock drifts.
While atomic clocks have been used in certain PL
applications where a ranging signal of high quality is
desired (Soon et al, 2003), most PLs typically use
inexpensive temperature compensated crystal oscillators
(TCXOs) to provide their reference frequency. Typical
TCXOs have a stability of around 1 part in 106, more
than six orders of magnitude worse than atomic
standards. Because of such insufficient stability in their
reference frequencies, PLs that operate asynchronously
cannot provide accurate stand-alone navigation. In such
cases double-differencing with a reference receiver is
commonly used to eliminate PL and receiver clock
biases. For a real-time solution this requires a wireless
data link between the user and reference receivers. This
adds operational constraints and affects performance,
depending on data link range, integrity and latency.
Furthermore the reference receiver must have visibility of
all PLs for accurate measurements. This may mean
having several reference receivers, for instance, in an
indoor PL network that spans several rooms. This of
course will add cost and complexity to the system. Also,
timing information is eliminated in the double-
differencing procedure. So, asynchronous PL operation
will be unsuitable for those applications in which precise
time transfer is important.
Asynchronous operation of PLs can also complicate the
pulsing technique often used for the mitigation of the
near-far effect. Consider the case of two PLs that are both
pulsed with a duty cycle of 10% over the C/A code
duration. If the PL pulsing scheme is asynchronous, the
transmission cycles of both PLs may overlap for
durations that vary according to the drift of both PL
clocks. This will impact on how well a receiver tracks
both PL signals. Since a receiver can only correlate for
10% of either PL’s transmission, overlapping of the two
signals can cause a significant increase in interference,
especially in a saturated receiver. In the worst case both
PLs will be mutually jamming each other.
An example of asynchronous PLs that are used to
augment GPS for an aircraft precision approach and
landing is given by Soon et al (2003). Kee et al (2000)
present a successful account of indoor positioning using
only asynchronous PLs. Cobb (1997) discusses a
synchronous pseudolite, known as a synchrolite. A
synchrolite acts as an electronic mirror to reflect GPS
satellite signals from a known point on the ground. A
synchrolite consists of a co-located GPS receiver and a
PL transmitter. The receiver is able to determine the
precise code phase and carrier frequency of the satellite
signal that it is tracking. This information is then used to
generate the PL transmissions that are synchronous to the
GPS signals. Synchrolite signal measurements are
typically differenced from satellite signal measurements
to eliminate spatially correlated errors.
Söderholm and Jokitalo (2002) present a synchronous PL
network for indoor positioning. The PLs in this network
are synchronized by a Master Control Station (MCS) that
tracks all PL signals using a reference receiver. Clock
corrections are communicated to all the PLs via hard
wired links. This centralized approach to synchronization
that mirrors the GPS OCS has several disadvantages.
First, the reference receiver must have visibility of all
PLs in the network. In the OCS this is achieved by
distributing monitoring stations around the world.
Similarly, the MCS needs to collect measurements from
as many reference receivers as are required to track all
PLs in the network. Second, the synchronization control
loop of the MCS must operate much more frequently than
its counterpart, the OCS. This is because PL clocks are
much less stable than the atomic standards used in GPS
satellites. Obviously, adding more PLs to the network
would also mean more computations for the MCS, and
this reduces the scalability of the system. Although hard
wired data links were presented, these could be
substituted by wireless links to eliminate the process of
laying communication cable during installation.
However, running such a high frequency synchronization
control loop over communication links to remote PLs will
undoubtedly add noise due to communication latency.
A decentralized approach to synchronization has been
demonstrated by Kee et al (2003) and Barnes et al (2003).
In a decentralized system, each PL has a co-located
receiver that tracks its own signal as well as a reference
signal. The reference signal may be the transmission of a
PL selected to be “master”, or that of a GPS satellite. PL
clock corrections can then be determined by taking
single-difference measurements of pseudo-range and
Integrated Carrier Phase (ICP) between the reference
signal and the local PL signal. Because the receiver is co-
located, control loop latencies are minimized and the PL
network is highly scalable; any new PLs can be added to
the network without being limited by data links or being
concerned about the visibility of PLs to a reference
receiver. Having a co-located receiver, however, also
adds to the cost and complexity of PLs.
In a decentralized system, various synchronization
topologies can be formed to suit the application. For
instance PLs are not required to track the “master” PL to
be synchronous to the network. A PL can be selected to
synchronize to another PL that is already synchronous to
the master PL. This propagation of time (i.e. PL
synchronization) has been demonstrated by Barnes et al
(2003). Clearly, propagation of PL synchronization
allows for easy expansion of PL networks. Consider an
indoor PL network for a warehouse that requires
expansion to cover outdoor adjacent delivery sites.
Additional PLs can be placed outside that only need to

158 Journal of Global Positioning Systems
track one indoor PL to achieve synchronization with the
existing network. Such a scheme is comparable to ad-hoc
wireless communication networks in which routing
information propagates to reflect changes in topology.
It must be said that propagation of synchronized PLs
requires further study. The process of synchronizing a
‘slave’ PL to a ‘master’ signal will undoubtedly produce
noise in the slave PL’s signal, the amount of which will
vary according to factors such as the bandwidth of the
synchronization loop. If another ‘slave’ PL is to then
synchronize to the signal of the first ‘slave’ PL, the
amount of noise in the transmission of the new PL will
increase. This can, of course, affect the accuracy of the
code and carrier measurements made from this signal.
2.1.4 Location errors and modelling
Near-far, multipath and synchronization are considered to
be the major issues that have to be addressed in PL
systems. However, there are other issues that arise, some
depending on the particular application of the PL system.
One issue concerns the accuracy of the specified location
at which a PL transmit antenna is mounted. A study of
the impact of PL location errors on positioning was
presented by Lee and Wang (2002). The observed effects
vary with the geometry between the PLs and the user
receivers. For a static receiver, the PL location error will
impose a bias in the PL measurements. For a non-static
receiver, single-differenced measurement errors can be up
to twice as large as the PL location error, depending on
the geometry. Thus mounting the PL antenna accurately
on a stable platform is necessary for high precision
applications.
Another issue that needs to be addressed for wide area PL
systems is tropospheric modelling. Signal propagation
delays through the troposphere vary with air pressure,
temperature and relative humidity. These delays can be
corrected for in the receiver by modelling the delay. The
tropospheric models used in GPS assume signals will
originate from 20,000 km in space, and the modelled
delays are highly dependant on satellite elevation.
Because PL networks aren’t as large, the same
tropospheric models as used in GPS won’t be as accurate.
Tropospheric delays can amount to 32 cm per km, so for
wide-area PL networks spanning several kilometres these
delays must be factored to obtain accurate measurements.
An overview of error modelling in PL applications is
presented by Dai et al (2001).
2.2 Issues specific to L1 C/A code pseudolites
2.2.1 Legal
The rational, equitable and efficient use of the
radiofrequency spectrum is coordinated by the
International Telecommunications Union (ITU). Member
states belonging to the ITU co-operate on agreements
concerning issues such as frequency allocation and the
registration of radio frequency assignments in order to
avoid harmful interference between radio stations. These
agreements are then implemented at a national level in
the form of laws and regulations that are applicable only
to the governed territory. As such, there may be slight
differences in the way frequency allocation is regulated
among the ITU’s member states.
The frequency band to which L1 belongs, 1559 – 1610
MHz, is reserved by the ITU for aeronautical
radionavigation and radionavigation-satellite (space-earth
and space-space) services. The aeronautical
radionavigation service refers to electronic aids such as
marker beacons and some aeronautical mobile
communications that form an integral part of aeronautical
radionavigation systems. The radionavigation-satellite
service refers to satellite based radio-determination
systems such as GPS and GLONASS.
The Federal Communication Commission (FCC) is the
government body in the United States of America that
regulates the use of the spectrum according to ITU
agreements. The FCC limits the use of the band 1559-
1626.5 MHz to airborne electronic aids for air navigation
and any associated land stations (47 CFR 87.475). Under
this regulation an L1 PL used in augmenting GPS in an
application such as aircraft approach and landing may be
considered as an aeronautical radionavigation aid. The
FCC states further that transmissions by aeronautical
radionavigation land stations must be limited to
aeronautical navigation (47 CFR 87.471). So, while
aeronautical applications of L1 PLs may comply with
existing FCC regulations, it seems that other applications
of L1 PLs are not permitted. However, the case may be
that L1 PLs deployed to provide air navigation may also
be useful for non-aeronautical purposes as well.
The FCC assigns the band 1215-1300 MHz primarily for
the military services and only permits limited secondary
use by other government agencies in support of
experimentation and research programs (47 CFR 2.106).
This regulation rules out any prospect of using L2 PLs,
and hence the benefits from using dual-frequency L1/L2
PLs, for any non-military application or for any extended
period of time. Dual-frequency PLs would allow on-the-
fly ambiguity resolution and also provide redundancy to
protect against multipath fades.

Kanli: Limitations of Pseudolite Systems Using Off-The-Shelf GPS Receivers 159
For researchers, however, FCC regulations permit
experimental radio service licences. Such a licence may
allow researchers to use any government or non-
government frequency, as designated by the FCC,
provided that the need for the frequency requested is fully
justified by the applicant (47 CFR 5). Licences are valid
for either 2 or 5 years with renewals awarded only upon
an adequate showing of need. It must be said that the
FCC will only permit an Experimental Radio Service
licence on the condition that harmful interference will not
be caused to any station operating in the frequency bands
allocated.
The Australian Communications Authority (ACA)
implements regulations based on ITU agreements in
Australia. As per the ITU allocation, the frequency band
1559 – 1610 MHz in Australia is also reserved for
aeronautical and radionavigation-satellite services.
However, with regards to all frequency bands, section
10(1) of the Australian Radiofrequency Spectrum plan
states that: ‘A frequency band may be used for an
unspecified service if the unspecified service uses the
frequency band to support a specified service’. Hence it
seems that L1/L2 PLs may be legally used to augment
GPS under existing laws. However, the use of L1/L2 PLs
in any such augmentation must not cause harmful
interference to the GPS service (section 10(5)). There is
no clarification on whether L1/L2 PLs can be used for
other purposes; however, the ACA also provides
experimental radio service licences for research purposes
with conditions similar to those imposed by the FCC.
A recent piece of legislation announced by the ACA
concerns the possession of radionavigation-satellite
service (RNSS) jamming devices (ACA, 2004). While
the operation of RNSS jamming devices was already
prohibited, this new legislation allows for the prosecution
of a person who supplies a RNSS jamming device or
possesses a RNSS jamming device for the purpose of
supply. The ACA defines a RNSS jamming device as “a
device that is designed to have an adverse effect on the
reception by RNSS receivers of RNSS
radiocommunications”. While L1/L2 PLs aren’t explicitly
designed for this purpose, they can easily be used to
prevent the reception of GPS signals. For instance the
popular IN200D PL by IntegriNautics, which has a peak
output power of +6 dBm, can jam GPS signals at
distances of up to 10 km. Even an L1-only PL can inhibit
the reception of L2 signals by jamming satellite C/A
codes on L1. This can happen because dual-frequency
receivers use the shorter C/A codes to acquire the week-
long P codes on L2. With such possibilities in mind, an
inquiry was put forward to the ACA as to whether L1/L2
PLs would be classified as RNSS jamming devices. If so,
this would certainly be of considerable concern to
researchers using/developing L1/L2 PL technology in
Australia. However, no reply was received by the time of
writing this paper.
From a rather brief look at radiofrequency spectrum
regulation in the USA and Australia, it is clear that
current legislation does not permit the widespread use of
L1 or L2 PLs. Commercially developed PLs often have
sections in their documentation that highlight this legal
issue. For instance the quick-start application note for the
IN200D by IntegriNautics states that users must check
the local laws and acquire any necessary licences and that
the licences be posted in public view during use. The
author is also aware of PL experiments performed for Air
Services Australia (ASA) during which a Notice to
Airmen (NOTAM) was issued to warn of possible local-
area interference to GPS signals. Such precautionary
measures are considered important by governing
authorities, even if interference to GPS signals may be
minimal. The legal issues involved in using L1/L2 PLs
are rarely discussed in PL research papers; the most
comprehensive overview found by the author is presented
by Cobb (1997).
2.2.2 Jamming GPS
GPS satellite signals must compete with each other and
with any other signals present in the L1/L2 frequency
bands. When competing amongst other satellite signals,
the auto-correlation properties of C/A codes allow a GPS
receiver to differentiate between the signals of the
satellites that are in view. C/A code cross-correlation
averages to about -30 dB when considering all Doppler
and time offsets; so there is about 30 dB of separation
between the satellite signals (the worst case cross-
correlation happens to be -21.6 dB).
It was stated earlier that GPS signals have a typical
strength of -160 dBW at the surface of the Earth. In
contrast, the power level of thermal noise at the antenna
of a GPS receiver is -205 dBW/Hz, which is -142 dBW
for the C/A code bandwidth of 2.046 MHz. The noise
figure of the amplifier immediately following the antenna
must also be considered. For typical OTS GPS receivers
this will increase the thermal noise to about -138 dB. This
means that GPS satellite signals are about 22 dB lower
than the noise floor. Since C/A codes are 1023 chips
long, a GPS receiver can achieve a processing gain of
about 30 dB if it integrates over a single C/A code period
(1ms). Typical OTS GPS receivers can integrate for up to
20ms which corresponds to a processing gain of 43 dB.
This processing gain allows the GPS receiver to raise the
satellite signals up above the noise floor. The result is a
post-correlation signal to noise ratio (SNR) of about 21
dB.
In the ideal case, where there are no other signals present
in the L1 band, GPS receivers won’t have any problem
acquiring or tracking unobstructed satellite signals.
Typical signal tracking loops of OTS GPS receivers can
acquire and track signals with a post-correlation strength

160 Journal of Global Positioning Systems
of 6 dB above the noise floor. This leaves a margin of
about 15 dB. As satellite signal strengths vary due to
factors such as satellite elevation and signal obstruction, a
receiver will still be able to track the signals provided the
theoretical 15 dB margin is not exceeded.
Forssell and Olsen (2003) performed a study on the
susceptibility of commercial GPS receivers to various
jamming signals. Modulated and unmodulated continuous
carrier waves as well as band-limited white noise centred
on the L1 frequency were subjected to three different
types of receivers. It was found that for the OTS receiver
used, GPS signals were unable to be tracked in the
presence of modulated interfering signals that were 36 dB
stronger than the GPS signals. White noise required being
about 53 dB stronger than GPS signals to have the same
effect. The study, however, did not look at spread
spectrum modulated interference signals, such as from a
PL. Although PL signals can be considered noise-like, PL
transmissions are modulated with C/A codes and as a
result cross-correlation effects with GPS signals will
become significant. Because of this PL signals are likely
to interfere at power levels much less than that of white
noise as reported by Forssell and Olsen.
Protection from interference to GPS signals can be
implemented at different levels. One level is at the
receiver’s antenna. A temporal (also spectral or notch)
filter at the antenna can provide some protection against
continuous wave CW interference, though such a filter
may be overwhelmed by interference from multiple
sources/jammers. With an antenna that has multiple
elements, interfering signals can be attenuated by
adjusting a weight applied to the outputs of each element.
Some techniques that use this approach are cancellation,
spatial temporal adaptive processing (STAP) and spatial
frequency adaptive processing (SFAP). Cancellation has
limited use, and although STAP and SFAP both can
provide 25-40 dB of jamming suppression, they do
require significant computation as well as high RF
dynamic range. Furthermore, the issue of phase
perturbations also has to be resolved.
Interference can also be suppressed internally within the
receiver. Narrow-width correlators, for example, provide
better immunity to interference than ordinary correlators.
This was shown in the experiments performed by Forssell
and Olsen (2003). A more simple yet effective method is
to increase the integration time. A typical OTS receiver is
limited to integrating for up to 20ms due to the presence
of the 50Hz navigation data bits, which cause 180 degree
phase shifts of the carrier. This limit may be overcome by
data wipe-off, or by using non-coherent integration. Data
wipe-off presumes prior knowledge of data bit values
from an external source, for which a data link is required
in real-time receivers. Non-coherent integration,
meanwhile, can lead to squaring loss and thus an increase
in noise. Nevertheless, although increased integration
times are achievable, it is done so at the expense of the
rate of measurement and navigation solution
computation. A better alternative in interference
cancellation involves coupling inertial navigation sensors
(INS) with the receiver tracking internals at varying
degrees. The most benefit is gained from tightly-coupled
schemes, though with this approach comes added cost
and complexity. Further details of interference
cancellation techniques mentioned here are presented by
Rounds (2004a and 2004b).
Although a variety of interference suppression techniques
are available, they are often limited to military or survey-
grade GPS receivers, which are generally quite
expensive. Typical OTS GPS receivers do not provide
sufficient immunity against interference to be of any
practical use with PL systems. To overcome this
limitation PL output can be set to a low enough power
level that avoids interference, though this can be a major
operational restriction depending on the type of
application. Another technique for minimizing
interference is to pulse the PL signals. Pulsing at low
duty cycles, which is often used to mitigate near-far
effects, allows PL signals to be stronger than GPS signals
without denying GPS to non-participating receivers.
To illustrate, consider a PL pulsing at a 10% duty cycle
over the C/A code period. This means that it is on for
100µs and off for 900µs. In this pulsing scheme, over a
C/A code period the GPS signals’ duration is 10 times
longer than that of the PL signal. This corresponds to a 10
dB difference in favour of the GPS signals. So, if a PL
signal whose power is 10 dB greater than GPS signals is
pulsed, to a GPS receiver the pulsed signal will appear to
have the same power level as the GPS signals. In this
case no interference will be caused, since the PL signal
power is averaged over the 1ms integration period. For
stronger PL signals, though, there will be interference
effects. A receiver tracking GPS signals will correlate
over only the GPS signals for 90% of the time. However,
for the 10% during which the PL is transmitting, the
receiver will also correlate over the PL’s in-band
transmissions. If the PL signal is significantly stronger
than the GPS signal, a reduced correlation gain in the
GPS signal will be observed during the PL pulse. The
effects of cross-correlation between the PL and GPS
signals will also become quite significant. These
combined factors result in a lower SNR for the GPS
signals as observed by a receiver. The longer the pulse
duration, the higher the reduction of SNR due to
increased interference.
Fig. 1 (Cobb, 1997) illustrates the relationship between
PL pulse duration and GPS signal SNR (S/N in the
figure), as observed by a typical GPS receiver. The PL
signal power is assumed to be at saturation level; which is
equal to the level of thermal noise 23 dB greater than the
GPS signals. The minimum SNR threshold at which a

Kanli: Limitations of Pseudolite Systems Using Off-The-Shelf GPS Receivers 161
receiver is able to track signals is indicated on the figure
to be about 6 dB. According to Cobb, this threshold is
exceeded for pulse duty cycles greater than 20% (curve
labelled ‘without blanking’). The minimum pulse duty
cycle required to track a PL signal is also determined
from the figure to be 10%. This means that at most two
such PLs outputting high power may operate within the
same area without denying GPS to non-participating
receivers.
Fig. 1 PL Pulse Duty Cycle Trade-off (Cobb, 1997).
2.2.3 Cross-correlation
The C/A codes used in GPS are taken from a family of
codes known as Gold codes. C/A codes were especially
chosen for their good multiple access properties over
their 1023 chip period. Their relatively short code length,
which is clocked at a 1.023 Mbps rate, also permits rapid
acquisition. Cross-correlation of C/A codes is dependant
on the delay offset between any given pair of codes. The
maximum cross-correlation of C/A codes is -24 dB, and
this worst-case is likely to occur 25% of the time. The
rest of the time cross-correlation is -60 dB; over time, the
average cross-correlation can be taken to be about -30
dB. The different C/A code used to modulate each
satellite signal isn’t the only major distinction between
any two signals. The orbit of GPS satellites impose a
Doppler frequency offset (up to +/- 6 kHz) on signals that
also affects cross-correlation. The worst case cross-
correlation over the expected Doppler offsets is -21.6 dB.
With GPS, the worst case cross-correlation is typically
not considered too important. This is because any
coincidence of Doppler offset and code offset that
manifests in a worst case is only transitory due to the
constant motion of the satellites. PLs on the other hand
are static devices. Therefore, a worst case combination of
frequency and code offset is unlikely to vary as much or
as often. Any variation will be entirely due to user
movements, which generate Doppler frequencies
considerably less than the motion of satellites. And, since
a C/A code chip at 1.023 Mbps is 300 m long, the
resulting changes in code phase due to typical user
movement are also considerably less.
In the example pulsing scheme of the previous section,
the PL signal pulsing was performed relative to the C/A
code period. Receivers will always integrate for at least
one C/A code period; typically over one period during
acquisition before increasing to as much as 20ms once bit
synchronization is achieved. So, the presented pulsing
scheme ensures that a PL signal pulse will always be
present during every integration period. If the PL pulse
duty cycles are not kept synchronous with the C/A code
period, then the case may arise where PL pulses are
absent from some integration periods but dominant in
other ones. This may cause some receivers to have
trouble with acquisition and tracking of PL signals
(Cobb, 1997).
For accurate and reliable tracking of C/A code modulated
signals it is important to correlate over the entire code
sequence. This is the only way to fully exploit the
orthogonal property of the C/A codes. Pulsing PL signals
thus presents a problem, since only a portion of the C/A
code is transmitted during the active-phase of a pulse
cycle. During the off-phase a receiver tracking the PL
will correlate against noise, GPS signals or the
transmission of another PL, depending on the
environment. In such a case, cross-correlation noise will
dominate. This effect places a hard limit on the minimum
pulse duty cycle. Some duty cycles as low as 7% have
been used while still reliably tracking C/A code PLs
(Cobb, 1997). More typical, though, is the RTCM
recommended duty cycle of 1/11 (~10%). This allows for
correlation with 93 chips of each PL signal’s code, and
allows a maximum of 11 PLs to be used at any given time
if one so desires (eg in an indoor PL only application).
With simple synchronous pulsing at a 1/11 duty cycle, a
PL will transmit the same 93 code chips every C/A code
period. This, however, does not utilize the whole code
sequence and may cause aliasing in the signal. A better
implementation would use a different code interval every
code period. For the 1/11 pulse duty cycle, the entire code
sequence can eventually be transmitted in 11 code
periods, which correspond to 11ms. Since typical GPS
receivers integrate for up to 20ms this method will allow
any receiver to correlate over the entire code sequence of
a pulsed PL signal.
If the C/A code is chipped at faster rates more code chips
can be transmitted during each pulse. For instance,
chipping at 10.23 Mbps will allow the complete
transmission of the C/A code sequence at the same pulse
rate presented earlier (100µs on, 900µs off). This offers a
huge advantage in correlation. A faster chipping rate also
increases the maximum number of PLs that can be
accommodated in a synchronous pulsing system.
Although some GPS receiver hardware may permit
modifications to allow for faster chipping rates, the same
cannot be said for all typical receivers. Another solution
to allow higher pulse duty cycles, proposed by Cobb
(1997), involves disabling the correlators during periods
of high cross-correlation interference. For instance a
correlator channel assigned to GPS signals can be

162 Journal of Global Positioning Systems
disabled during PL active pulse cycles. The effect of this
is to prevent correlation with the stronger PL signals, thus
eliminating cross-correlation influences. Inversely,
correlators assigned to PL signals can be disabled during
pulse-off cycles to eliminate cross-correlation noise in the
same manner. The expected benefit of this technique,
referred to as correlator blanking, is depicted in Fig. 1.
The satellite signal SNR curve (labelled ‘with blanking’
in the figure) shows that a higher pulse duty cycle of up
to 60% can be achieved. Cobb also briefly discusses an
implementation of correlator blanking, stating that a
saturation detector at the front-end could serve as an
indication of PL pulses. Correlators can then be enabled
or disabled accordingly. Unfortunately, this requires
significant modification to a typical receiver, but may be
easily incorporated into newer receiver designs. A
problem with this implementation is that it does not
address how two pulses from different PLs can be
distinguished. All correlators assigned to a PL will be
enabled during saturation. If there are two PLs the
receiver will correlate over pulses from both devices,
increasing cross-correlation noise. Also, as the distance to
a PL increases the PL signal will cease to saturate the
receiver. In this case, the saturation detector will not
respond to the PL signal, leaving the correlator disabled
during the PL pulse. Rather than helping, correlator
blanking can prevent the acquisition and tracking of a PL
signal; not correlating at all is worse than correlating
against interference for 90% of the time. A further
extension to this tragic case is when a pulse from a near
PL enables the correlator assigned to the far PL, this time
leaving the receiver to correlate against the wrong signal.
2.2.4 C/A code chip rate in high multipath
Multipath is a major concern for PL systems that operate
indoors and in other cluttered environments. These
environments generate multipath signals that may
outnumber the direct signal by many orders of magnitude,
and each may have different delays, strengths and
polarizations. According to a simple analysis, assuming
the multipath signal is never stronger than the direct
signal, the peak multipath error can equal one half of a
chip length (Braasch, 1996). This evaluates to nearly 150
m at the C/A code chipping rate. The actual error can be a
lot worse when considering the myriad of different
reflections in, say, a warehouse or urban canyon. It is
therefore unlikely that a chipping rate of 1.023 Mbps can
provide an acceptable level of multipath rejection in high
multipath environments. Faster chipping rates however,
will have reduced peak multipath errors. For example, by
the same analysis, the 10.23 Mbps rate used for P codes
has a peak multipath error of only 15 m.
2.2.5 Receiver front-end saturation
A typical GPS receiver front-end consists of an antenna,
followed immediately by a low noise amplifier (LNA),
then a filter, one or more stages of down-conversion, and
finally an analog-to-digital converter (ADC). An
automatic gain control (AGC) may also be implemented
to keep the input signal within the dynamic range of the
front-end. (Dynamic range refers to the range of signal
powers that a component (or system of components) can
process without generating distortion.)
Recall that the GPS signals at the Earth’s surface are
extremely weak. When unobstructed, these signals are
approximately of equal strength, except for minor
variations due to changes in satellite elevation. GPS
receiver manufacturers take advantage of these
characteristics of GPS signals to simplify the design of
their receivers. One feature that is common of typical
GPS receivers is the low dynamic range of the front-end.
Since GPS signals are buried 22 dB below the noise floor
at a relatively constant level, it makes little economic
sense to design an RF front-end that compensates for
large signal power deviations when only slight changes in
noise level due to changes in temperature is expected.
A receiver will see only thermal noise when no
transmissions other than the GPS signals are present in
the L bands. In the presence of strong interference, like a
PL signal 30 dB above noise, the receiver’s AGC will
reduce gain to compensate for increased signal power.
The AGC tries to scale the input signal, which is expected
to be thermal noise, over the full dynamic range of the
ADC. The result, though, is that the thermal noise and
GPS signals are attenuated along with the interfering
signal. If the attenuation exceeds about 15 dB then the
receiver will be unable to track GPS signals.
Furthermore, if the reduction of gain is insufficient, the
interfering PL signal will also be clipped at the
extremities of the ADC’s range. The receiver then is said
to be in saturation.
In the presence of such interference, the RF components
used in a typical GPS receiver can be operating outside of
their dynamic range. In effect, they are operating beyond
the limits of their specification, in which their function
may become non-linear. One important component, the
mixer used in down-conversion, should always be
operating in its linear range. Otherwise, if the input power
exceeds its threshold, compression of the received signal
can result that would generate harmonics which cause
distortion. Sustained excessive interference may even
cause damage to some components.
Typical GPS receivers in saturation will perform quite
poorly simply because they are not designed to operate in
this condition. In these receivers, interference and cross-
correlation effects are magnified. Detection thresholds
must be increased to avoid locking onto false peaks.

Kanli: Limitations of Pseudolite Systems Using Off-The-Shelf GPS Receivers 163
Furthermore, a receiver can only track one saturating PL
signal at a time, and only if it significantly stronger than
the other PL signals. Two saturating PL signals of the
same strength will interfere with each other so that
neither can be accurately tracked. This has implications
for pulsed PL networks. As mentioned in section 2.1.3,
the active-phase of a pulse cycle from multiple PLs may
overlap in time due to PL clock drift. This will decrease
PL signal SNR, due to partial jamming, and may also
cause tracking errors or prevent acquisition. The pulsing
scheme of multiple PLs should be synchronized so that
duty cycles do not coincide. In wide-area deployment of
PLs attention must also be given to propagation delays
that may cause some overlapping of adjacent
synchronous PL pulses. A more thorough discussion of
the effects of GPS receiver saturation is presented by
Cobb (1997).
3 Impact of L1 C/A code issues on pseudolite
applications
This section will briefly outline how the presented issues
may impact on PL usage for several different
applications.
3.1 Aircraft approach and landing
Legal: L1 C/A code PLs used for aiding aircraft
navigation appear to be legal under FCC and ACA
regulations.
Jamming/Saturation: Radionavigation systems for
aircraft are subject to stringent performance requirements.
Therefore, any factors that limit the accuracy and
reliability of these systems are of great concern. PLs must
provide adequate coverage to enable aircraft to acquire
and track their signals well before the landing approach.
This involves distances of up to 15 km. In order to
provide this coverage, PLs must transmit at high power,
meaning receivers at closer distances will be saturated.
The saturation of receivers by L1 C/A code PL signals
limit the number of PLs that can be deployed around an
airfield to about 2. Users in this system will be denied
more PLs that could offer better geometry and provide
extra ranging signals for reliability. The jamming effect
of high power PL signals combined with saturation
effects also degrades the performance of participating and
non-participating receivers.
3.2 Integrated GPS/pseudolite positioning
Legal: According to current FCC regulations, operating
an L1 PL is illegal without an experimental radio station
(or other) licence. It is unclear as to whether any changes
in regulations will be made. For research work which
typically involves limited intervals of PL transmission
this legal obstacle may be considered minor. However,
from a commercial perspective this is a major hurdle that
stands in the way of implementing a round-the-clock L1
PL service which is integrated with GPS. Furthermore,
under current ACA regulations designers and suppliers of
L1 PLs, which may be considered as RNSS jamming
devices, are liable to prosecution and heavy fines.
Jamming/Saturation: From a consumer’s perspective,
those who have bought and paid for GPS receivers that
are either unable to track PL signals or are unable to
handle even minimal interference from PL signals, are
unlikely to support the deployment of L1 C/A code PLs
in their area. PLs may instead be configured to transmit at
low power, although this requires their deployment in
increased numbers to provide the required coverage.
3.3 Indoor pseudolite-only systems
Legal: Operating L1 PLs indoors without a licence, even
at low power levels, is still considered illegal.
Jamming/Saturation: It is difficult to maintain constant
signal levels in cluttered indoor environments, even if the
near-far effect was eliminated. PL signals will be
received at power levels varying over a very wide range,
especially if signals are required to propagate through
walls and ceilings. A weak signal from an adjacent room
will most likely be jammed by a closer PL. Also, current
OTS receivers with their low dynamic range will exhibit
difficulties in coping with the high range of PL signal
strengths. Receivers may frequently move in and out of
saturation. GPS signals indoors are so heavily attenuated
that they are unusable by most receivers. From this
perspective, the jamming of indoor GPS signals by L1
PLs could be forgiven. However, new Assisted-GPS
(AGPS) techniques have been developed that enable
special GPS receiver designs to acquire and track weak
indoor signals. In this case, jamming GPS signals is still
an issue for indoor systems.
Multipath: By far the greatest limitation of L1 C/A code
PLs in indoor applications is the low multipath mitigation
offered by its signal structure. The chipping rate of the
C/A code is just too low, leaving the receivers vulnerable
to high multipath range errors.
4 Future directions
Research into the use of PLs for many different
applications has been conducted for over two decades.
Throughout this time almost all of the experimentation
has involved PLs using the civilian signal structure of
GPS. The primary reason for this was that it enabled the

164 Journal of Global Positioning Systems
use of existing receiver hardware with only minor
firmware modifications. Innovative methods of
incorporating PLs into navigation applications were
developed. Promise in this new technology led several
companies providing positioning solutions to develop and
sell L1 C/A code PLs. However, to date the use of PLs
has been limited to the realm of research or to highly
specialized applications. The author believes that this is
due to the various issues which limit the effectiveness of
L1 C/A code PLs as discussed in the previous sections.
With the development of Galileo and the modernization
of GPS, some of the issues in this paper will be
addressed. One is the faster chipping rate on civilian
signals that will reduce maximum multipath errors; and
the other is the two civilian frequencies that provide a
means for ambiguity resolution. When both these systems
are operational, PLs augmenting Galileo and modernized
GPS on the new civilian signals will be able to provide
some performance gains over the current PLs augmenting
GPS. Unfortunately, no benefit will be gained with
regards to jamming, interference or receiver saturation.
Furthermore, if a provision for the use of PLs to augment
the new signals is not defined, the same legal obstacles
will remain in place, limiting their widespread use.
A current trend in PL research and development is the
movement away from preserving backward compatibility
with existing GPS receivers. Most of the ideas are not
new, but the driving force behind them is a recent
realization that a signal structure used for global
positioning cannot always be used for local positioning,
especially indoors. One of the ideas being implemented is
multi-frequency transmissions in the Industrial Scientific
and Medical (ISM) band (Zimmerman et al, 2000).
Operating in the ISM band requires no licence, as long as
the maximum power output is limited to less than 1 Watt
(FCC-47CFR15.247). The major advantage gained is the
legal freedom to transmit on multiple bands, including
915 MHz, 2.45 GHz and 5.8 GHz. Receivers can exploit
the frequency diversity to resolve carrier cycle
ambiguities. On the other hand, there may be interference
issues with other devices that also use the ISM band, such
as cordless phones and 802.11b/g.
Locata Corporation, an Australian company based in
Canberra, is currently developing PL technology for
operation in the ISM band. Locata Corporation is
building on the success of its synchronous L1 prototype
PLs that have proven accurate performance in an indoor,
high multipath environment (Barnes et al, 2004). A
Locata PL, called a LocataLite, is an intelligent
transceiver that transmits on dual frequencies in the ISM
band. A LocataLite is able to synchronize to the
transmissions of other LocataLites, forming a
synchronous positioning network called a LocataNet. A
receiver device, called a Locata, can determine its
position within the LocataNet using both code and carrier
phase measurements. The main advantage offered by a
LocataNet is its synchronous signals, which allow
standalone navigation and precise time transfer.
Novariant (formerly IntegriNautics) has also developed
an off-frequency PL system called a TerraliteTM
(Novariant, 2004). A TerraliteTM transmits a proprietary
signal called XPS. At the time of writing, little public
information was available on the frequency or signal
structure of the XPS signal, though the requirement for a
licence was announced. Novariant also provides a tri-
frequency receiver that is able to track both L1 and L2
GPS signals in addition to the XPS signal. Novariant is
targeting specialized heavy industries, such as mining,
where GPS is one of the primary positioning systems
used. In such industries considerable investment has been
made into GPS receivers and even into L1 C/A code PLs
in order to address their positioning needs. However, in
light of the issues discussed in the previous sections, this
can only provide limited performance. The TerraliteTM
XPS signal offers an additional signal designed to
overcome the restrictions of L1 PLs that are currently
used. Furthermore, the ability of the XPS receiver to also
receive both L1 and L2 signals means that a client’s
existing GPS infrastructure can still be utilized.
Multi-frequency PLs can complicate the design of the
front-end of receivers that are intended to process them.
For example the Novariant XPS receiver, which must
acquire L1, L2 and a third frequency. The number of
components required will increase; especially if separate
RF paths are to be used for each frequency. This can add
significant cost to receivers. Currently, research is being
performed on reducing the need for a complicated front-
end by using direct sampling instead (Psiaki et al, 2003).
Another development in receiver technology is the
software correlator receiver. These receivers perform all
signal processing tasks after digital conversion, including
correlation, within a multi-purpose processor. Software
receivers can offer flexibility in adapting to the various
signal modulation plans of future PL systems. However,
significant processing power will be required for
processing signals with high chipping rates. Pany and
Eissfeller (2004) address this performance issue by using
sub-Nyquist sample rates.
In the past, jamming of GPS signals has been primarily a
concern for the military. However, beyond military
concerns are civilian vulnerabilities related to critical
infrastructure. Cell phone networks, commercial fishing,
transportation, emergency response, air navigation and air
traffic control; all have developed a reliance on GPS for
timing and navigation. The list of dependencies will grow
as accurate position information becomes more tightly
integrated into business practices. This is expected to spur
the development of more robust receivers to address
some of the issues in jamming and interference.

Kanli: Limitations of Pseudolite Systems Using Off-The-Shelf GPS Receivers 165
For indoor positioning, the high multipath environment
necessitates the use of more robust signal structures. An
example is Locata Corporation’s synchronously pulsed
signals with fast chipping rates. Also, their use of
multiple frequencies provides redundancy to protect
against multipath fades. Another is presented by Progri et
al (2004), who propose a very unique indoor positioning
system using ultra-wideband (UWB) signals. This system
uses orthogonal frequency division multiplexing (OFDM)
with frequency division multiple access (FDMA) that
minimizes cross-channel interference. The UWB signals
also offer good protection from multipath fades.
Other techniques for indoor positioning have been
proposed. One is Siemen’s Local Positioning Radar
(LPR) system (Siemens, 2004). This system comprises of
transponders that are deployed around the area of interest.
A base station transmits pulses which are received and
retransmitted by each transponder. The base station will
receive each retransmitted signal at different times
according to its distance from each transponder, hence
providing range information. Another is a UWB time-of-
arrival (TOA) technique (Fontana and Gunderson, 2002).
In this system, passive receivers in a chained network
detect the pulsed transmissions from active UWB tags.
TOA data from the receivers are then processed to
determine the position of each active UWB tag.
Assisted GPS (AGPS) techniques provided by Global
Locate, SiRF and Q-Comm use massively parallel
correlator architectures to detect the heavily attenuated
GPS signals indoors (Global Locate, 2004). To do this, a
data link provides known navigation messages to the
AGPS receiver, enabling it to integrate for longer periods.
However, considerable infrastructure is needed to provide
navigation messages to all users.
Rosum Corporation proposes the use of digital television
signals to augment GPS (Rosum, 2004). TV signals use a
variety of VHF and UHF bands, hence giving protection
against multipath fades. Their signals are also stronger
than GPS signals by about 40 dB. However, this
technique also requires significant infrastructure to
monitor channel stability and timing information. Also,
currently receivers using this technology are unable to
calculate their own position; this is performed at a
location server and sent back to a receiver because
current TV signals are not synchronous to a common
clock.
In some of the techniques above, the position of a
location device is determined by the system, not by the
device itself. This may be desirable for applications such
as centralized asset management, but would be unsuitable
for other applications such as autonomous robotics or
guidance and control. Furthermore, the reliance on data
links for navigation messages or timing information may
be a hindrance. For these and other reasons, the author
believes that it is unlikely for any single positioning
system to be the answer to every application’s needs.
What the author does believe is important, though, is that
the users of positioning technology should have full
control over their signals. Only then can a positioning
system be optimized to best suit one’s needs.
5 Conclusions
L1 C/A code PLs simplify experimentation with PL
technology. They allow the convenient use of existing
OTS GPS receivers. However, this practicality has
disadvantages. One involves the legal issues of
transmitting on L1, a band protected globally by
legislation. L1 PLs can also jam GPS signals, denying
GPS to non-participating receivers. The C/A code
chipping rate used by these PLs also does not provide
sufficient protection against multipath. Typical OTS GPS
receivers are easily saturated by strong PL signals; this
limits the accuracy and reliability of the PL system.
These factors add further complications to general PL
issues such as near-far, multipath and synchronization. A
current trend in PL development serves to address these
complications by moving off the L1 frequency and using
more robust signal structures.
7 Acknowledgements
The author would like to thank Dr. Jimmy Lamance,
Dave Small, Dr. Joel Barnes, and Dr Dorota Grejner-
Brzezinska for their valuable input and assistance.
References
ACA (2004): Media Release No. 73, 1 September,
http://www.aca.gov.au/aca_home/media_releases/media_e
nquiries/2004/04-73.htm
Barnes J.; Rizos C.; Wang J.; Small D.; Voigt G.; Gambale N.
(2003): High precision indoor and outdoor positioning
using LocataNet, 2003 Int. Symp. on GPS/GNSS, Tokyo,
Japan, 15-18 November, 9-18
Barnes J.; Rizos C.; Kanli M.; Small D.; Voigt G.; Gambale N.;
Lamance J.; Nunan T.; Reid C. (2004): Indoor industrial
machine guidance using Locata: A pilot study at
BlueScope Steel, Proceedings of the 60th Annual Meeting
of the U.S. Inst. Of Navigation, Dayton, Ohio, 7-9 June,
533-540.
Braasch M.S. (1996): Multipath Effects, Global Positioning
System: Theory and Applications I, AIAA, 547-568.
Cobb H.S. (1997): GPS pseudolites: Theory, design, and
applications, PhD Dissertation, Stanford University.
Dai L.; Wang J.; Tsujii T.; Rizos C. (2001): Pseudolite
Applications in Positioning and Navigation: Modelling
and Geometric Analysis, Int. Symp. on Kinematic Systems

166 Journal of Global Positioning Systems
in Geodesy, Geomatics & Navigation, Banff, Canada, 5-8
June, 482-489.
Fontana R.J.; Gunderson S.J.; (2002): Ultra-WideBand
Precision Asset Location System, IEEE Conference on
Ultra Wideband Systems and Technologies, May 2002,
Baltimore, MD.
Forssell B.; Olsen T.B. (2003): Jamming GPS: Susceptibility
of Some Civil GPS Receivers, GPS World, Vol. 14, No. 1,
January, 54-58.
Global Locate (2004): Global Locate: Home,
http://www.globallocate.com
Irsigler M.; Eissfeller B. (2003): Comparison of Multipath
Mitigation Techniques with Consideration of Future
Signal Structures, Proceedings of the 16th Int. Tech.
Meeting of the Satellite Division of the U.S. Inst. of
Navigation, Portland, Oregon, 9-12 September, 2584-2592.
Klein D.; Parkinson B.W. (1986): The use of pseudo-satellites
for improving GPS performance, Global Positioning
System (red book), Vol III, Institute of Navigation, 135-
146.
Kee C.; Jun H.; Yun D. (2000): Development of Indoor
Navigation System using Asynchronous Pseudolites,
Proceedings of US Institute of Navigation GPS-2000, Salt
Lake City, Utah, 19-22 September, 1038-1045.
Kunysz W. (2003): A Three Dimensional Choke Ring Ground
Plane Antenna, Proceedings of the 16th Int. Tech.
Meeting of the Satellite Division of the U.S. Inst. of
Navigation, Portland, Oregon, 9-12 September, 1883-1888.
Wang J.; Lee H.K. (2002): Impact of Pseudolite Location
Errors on Positioning, Geomatics Research Australasia,
77, 81-94.
Madhani P.H.; Axelrad P. (2001): Mitigation of the Near-Far
Problem by Successive Interference Cancellation,
Proceedings of the 14th Int. Tech. Meeting of the Satellite
Division of the U.S. Inst. of Navigation, Salt Lake City,
Utah, 11-14 September, 148-154.
Misra P.; Enge P. (2001): Global Positioning System: Signals,
Measurements, and Performance, Ganga-Jamuna Press,
Lincoln, Massachusetts, 2001.
Novariant (2004): Novariant: Mining – Products,
http://www.novariant.com/mining/products/index.cfm
Pany T.; Eissfeller B. (2004): Code and Phase Tracking of
Generic PRN Signals with Sub-Nyquist Sample Rates,
Journal of the Institute of Navigation, Vol. 51, No. 2,
Summer 2004, 143-159.
Progri I.F.; Michalson W.R.; Cyganski D. (2004): An
OFDM/FDMA Indoor Geolocation System, Journal of the
Institute of Navigation, Vol. 51, No. 2, Summer 2004, 133-
142.
Psiaki M.L.; Powell S.P.; Jung H.; Kintner Jr. P.M. (2003):
Design and Practical Implementation of Multi-
Frequency RF Front Ends Using Direct RF Sampling,
Proceedings of the 16th Int. Tech. Meeting of the Satellite
Division of the U.S. Inst. of Navigation, Portland, Oregon,
9-12 September, 90-102.
Rosum (2004): Rosum Corporation. Reliable Location. Inside
and Out, http://www.rosum.com
Rounds S (2004a): Jamming protection of GPS receivers; Part
I: Receiver Enhancements, GPS World, Vol. 15, No. 1,
January.
Rounds S (2004b): Jamming protection of GPS receivers; Part
II: Antenna Enhancements, GPS World, Vol. 15, No. 2,
February, 38 - 45.
Siemens (2004): Methods and Apparatus for Determining a
Location of a Device, United States Patent Application
Publication No. US 2004/0157625 A1.
Söderholm S.; Juhola T.; Saarnimo T.; Karttunen V. (2001):
Indoor Navigation Using a GPS Receiver, Proceedings of
the 14th Int. Tech. Meeting of the Satellite Division of the
U.S. Inst. of Navigation, Salt Lake City, Utah, 11-14
September, 1479-1486.
Söderholm S.; Jokitalo T. (2002): Synchronized Pseudolites -
The Key to Indoor Navigation, Proceedings of the 16th
Int. Tech. Meeting of the Satellite Division of the U.S. Inst.
of Navigation, Portland, Oregon, 9-12 September, 226-
230.
Soon B.H.K.; Poh E.K.; Barnes J.; Zhang J.; Lee H.K.; Lee
H.K.; Rizos C. (2003): Flight test results of precision
approach and landing augmented by airport pseudolites,
Proceedings of the 16th Int. Tech. Meeting of the Satellite
Division of the U.S. Institute of Navigation, Portland,
Oregon, 9-12 September, 2318-2325.
Stolk K.; Brown A. (2003): Phase Center Calibration and
Multipath Test Results of a Digital Beam-Steered
Antenna Array, Proceedings of the 16th Int. Tech.
Meeting of the Satellite Division of the U.S. Institute of
Navigation, Portland, Oregon, 9-12 September, 1889-1897.
Van Dierendonck A.; Fenton P.; Ford T. (1992): Theory and
Performance of Narrow Correlator Spacing in a GPS
Receiver, Navigation, Volume 39, Number 3, 265-283.
Wang J (2002) Pseudolite Applications in Positioning and
Navigation: Progress and problems, Journal of Global
Positioning Systems, 1(1): 48-56.
Weiss J.P.; Anderson S.; Axelrad P. (2003): Multipath
Modeling and Test Results for JPALS Ground Station
Receivers, Proceedings of the 16th Int. Tech. Meeting of
the Satellite Division of the U.S. Inst. of Navigation,
Portland, Oregon, 9-12 September, 1801-1811.
Yun D.; Kee C. (2002): Centimeter Accuracy Stand-alone
Indoor Navigation System By Synchronized Pseudolite
Constellation, Proceedings of the 16th Int. Tech. Meeting
of the Satellite Division of the U.S. Inst. of Navigation,
Portland, Oregon, 9-12 September, 213-225.
Zimmerman K.R.; Cohen C.E.; Lawrence D.G.; Montgomery
P.Y.; Cobb H.S.; Melto W.C. (2000): Multi-Frequency
Pseudolites for Instantaneous Carrier Ambiguity
Resolution, Proceedings of US Institute of Navigation
GPS-2000, Salt Lake City, Utah, 19-22 September, 1024-
1030.