Journal of Global Positioning Systems (2007)
Vol.6, No.2:158-165
Achieving Centimetre-level Positioning Accuracy in Urban Canyons
with Locata Technology
Jean-Philippe Montillet, X. Meng, G. W. Roberts, A. Taha, C. Hancock, Oluropo Ogundipe
Institute of Engineering Surveying and Space Geodesy, the University of Nottingham, University Park, Nottingham, NG7 2RD, United Kingdom
Joel Barnes
School of Surveying and Spatial Information Sy s tem s, University of New South Wales, Sydney, NSW 2052, Australia
Abstract. In 2005 The UK Department for Trade and
Industry (DTI) commenced funding a project called
Visualising Integrated Information on Buried Assets to
Reduce Streetworks (VISTA). The project aims to
precisely map buried assets (gas pipes, telecom cables,
etc) and increase the efficiency of the process in
challenging environments such as in urban canyons,
where GPS fails to work or is not reliable enough to get a
precise position. In this context the Institute of
Engineering Surveying and Space Geodesy (IESSG) at
the University of Nottingham purchased, at the beginning
of 2007, a terrestrial network positioning system called
Locata technology. This technology is developed by
Locata Corporation Pty Ltd from Australia. Over the last
five months researchers have carried out experiments
with this new technology on the main campus of the
University of Nottingham. The preliminary results show
that LocataLites are a suitable technology to solve the
positioning problems for the VISTA project. The overall
accuracy is at the centimetre level for all points surveyed.
Moreover, we underline in this paper the reliability and
the flexibility of this new technology.
Keywords: Multipath, RTK-GPS, Locata, LocataLites,
LocataNet, Positioning, Navigation.
1 Locata Technology
Currently there is an urgent need to map 4 million
kilometres of underground cables and pipes in the UK
alone — a combination of water, sewer, gas, electricity
and drainage infrastructure. Many of today’s buried water
and sewerage assets were laid during Victorian ti mes (up
to 200 years ago) when accurate records of the location
and depth of each pipe were not kept. Nowadays,
technologies like ground probing radar (GPR) make it
possible to detect the pipes without digging a hole.
Regardless, every time a company does dig there is a high
probability of hitting one of these pipes, causing severe
disruption for workers and customers. Hitting a live
power cable can, in some cases, prove fatal for workers
(Parker, 2006).
As underground asset location is such a hit and miss
affair, the UK’s Department for Trade and Industry,
along with a group of industry partners, funded in mid-
2005 a £2.4 million project called “Visualizin g Integrated
Information on Buried Assets to Reduce Streetworks”
(VISTA). The aim of the project is to meet this
recognized need to precisely map buried assets in urban
areas (Roberts et al., 2006).
Researchers at the University of Nottingham are working
on the VISTA project in collaboration with colleagues at
Leeds University, and with other industrial partners such
as UK Water Industry Research.
The work at Leeds University focuses, on one hand, on
gathering records from the various utilities, digitizing
their maps and building a database (Boukhelifa & Duke,
2007). On the other hand, the academics at the University
of Nottingham are working on how satellite technology
could be used to access information on where utilities are
buried, and so provide an accurate location of the assets
buried in the ground. Th e project aims to map any buried
assets within the centimeter level of accuracy. With the
geomatic products available in today’s market, this
accuracy is only feasible by using Real Time Kinematic
GPS (RTK-GP S) po si t i oni n g t echn ol o gy .
However, to work effectively and obtain accurate
coordinates, RTK-GPS receivers must be able to track a
minimum of five (and preferably more) well-distributed
satellites orbiting the earth. This can be a serious issue in
built-up areas and particularly in urban canyons.

Montillet: Achieving centimetre-level positioning accuracy in urban canyons with Locata technology 159
Furthermore, the performance of the GPS degrades
quickly in the high multipath environments present in
cities. Thus, GPS technology is not always available, not
reliable and not accurate enough in the dense multipath
environments in urban areas.
In previous studies, researchers at the IESSG have been
working on the integration of GPS and GSM-phone
signals, and have developed simulator tools to assist the
research. This work is comprehensively explained in
(Montillet et al., 2007). Although the results are
promising, the IESSG is now investigating Locata
technology (using LocataLites). Locata technology is a
new positioning technology developed by Locata
Corporation of Australia which uses a network of
ground-based transceivers that cover a specific area
(Barnes et al., 2006). In February 2007, the IESSG
(University of Nottingham) purch ased a Locata system in
order to demonstrate the technology proof-of-concept.
The goal of this present work is to demonstrate the
accuracy achieved so far with Locata technology, and to
compare the results obtained in different environments.
Section 2 starts with a brief history of the development of
Locata technology. It is followed with an introduction to
the main features of this ground-breaking system
(overview, signal and network synchronization). The
experimental setups are explained in Section 3, and the
results are discussed. The paper ends with the conclusions
supported by the experiments, and envisaged future work.
2 Locata Technology
2.1 History
Since the earliest day of GPS po sitioning (1978), ground-
based transmitters have been under development to
compliment satellite constellations. These ground-based
transmitters were called Pseudo-satellites or Pseudolites.
They have been used to test GPS system elements and
enhance GPS in certain applications by providing better
accuracy, integrity or availability through the use of
Pseudolite signals in addition to the GPS signals.
Pseudolites were also a promising technology for
providing positioning in high-multipath environments
where GPS signals are generally unavailable, severely
attenuated, or of poor quality. Thus they presented the
prospect of being useful for both indoor and outdoor
positioning app lications by tran smitting a GPS-like signal
(Cobb, 1997). Pseudolites work in an unsynchronized
mode and double differencing must be used to eliminate
the Pseudolites’ and receivers’ clock biases. Growing
interest in the mid-1990s foresaw Pseudolite technology
as the next “big thing”. Since then numerous Pseudolite
applications have been attempted: Local Area
Augmentation System (LAAS), plane landing, bridge
deformation monitoring, open pit-mining, reducing
streetworks (Cosser, 2004), (Misra & Enge, 2001),
(Kanli, 2005), (Van Dierendonck, 1997), (Roberts et al.,
2006).
However, there are many fundamental issues that limit
the effectiveness of a Pseudolite system using C/A code
on L1/L2. They include the illegality of transmitting on
L1/L2, cross-correlation between Pseudolites and GPS
signals (GPS jamming), saturation of GPS receiver front-
ends, and the limited multipath mitigation offered by C/A
codes. When combined with other problems inherent to
all Pseudolite systems such as near-far, multipath, and
synchronisation, the issues in using L1/L2 C/A code
Pseudolite systems further complicates the design and
deployment of such systems, and places limits on any
operational effectiveness (Kan li, 2005). If Pseudolites can
be synchronised in some manner, stand-alone p ositioning
can be achieved without base station data (and without
the need for a radio modem data link) (Barnes et al.,
2003). A couple of years ago, attempts to synchronise
Pseudolites resulted in position solutions that are up to
six times worse in comparison to an unsynchronised
approach using double-differencing (Yun et al., 2002).
Recently, two positioning solutions have emerged from
the development of Pseudolite technology: Terralites
(Novariant, 2005) and LocataLites . Locata is welcomed
as a new break-through in the ground-based positioning
world. The technology cons ists of a network (LocataNet)
of time-synchronised transceivers (LocataLites) allowing
point positioning of a rover with centimetre accuracy
(using carrier-phase). (Barnes et al., 2006).
2.2 Locata Technology a t a Glance
As explained below, Locata technology is built on new
proprietary synchron ization technology which over comes
the challenges presented when trying to use ground-based
transmitters. LocataLite transceivers transmit and receive
a signal modulated in the same way as the GPS code,
allowing a rover to trilaterate to calculate its position. A
LocataNet is a network of LocataLites. It consists of two
kinds of devices: the LocataLites and the Locata receiver
(or rover). The LocataLite transmitter generates a carrier-
phase signal modulated with a proprietary ranging code
in the 2.4GHz Industrial Scientific Medical (ISM) band
(Prasad, 1998). At the time of writing, the LocataLites
can transmit two positioning signals at the same
frequency with different Pseudo Random Noise (PRN)
ranging codes from the two transmit antennas. A third
antenna is used by the LocataLite to receive signals. Fig.
1 shows a LocataLite installation used for the
experiments in the car park of the IESSG.

160 Journal of Global Positioning Systems
Fig. 1 A LocataLite with a GPS antenna on top of it
Notice that there is a GPS antenna at the very top of the
mast. This GPS receiver is used only to give the exact
initial location of the LocataLites’ antennas (Tx1 and
Tx2). The waterproof metallic box (blue circle) protects
the LocataLites’ hardware. In the network calibration
process, the positions of all transmitting antennas are
monitored precisely, and are registered in the memory
card on board each LocataLite and the rover (Montillet,
2007).
Fig. 2 Rover and LocataLite
Moreover, it is common knowledge that deploying an
antenna array instead of a single antenna at the
transmitter side helps to protect the radio signals from
fading effects at the receiver side (Proakis, 2000). As a
rover receives two signals, each distorted by different
propagation paths after being transmitted from a
LocataLite, it can compare the Signal-to-Noise-Ratio
(SNR) and other multi-path mitigation qualities of the
two incoming signals. The rover can then detect if the
signal transmitted from one antenna is in a deeper multi-
path fading zone than the other one, and can modify the
processing of the signals in the trilateration process.
Finally, the receiver chipset and the transmitters share th e
same clock, which is a cheap temperature-compensated
crystal oscillator (TCXO) according to (Barnes et al.
2003b).
2.3 Locata Signal
At the time of writing this paper, the LocataLite
transceivers transmit single frequency ranging signals
(pseudo-range and carrier phase measurements) in the 2.4
GHz license free band. The carrier-phase equation can be
expressed as Eq. 1:
(
)
jj
AAtropA
j
ANTcR φ
ετ
λ
φ ++∂⋅++= 1 (1)
where A R is the geometrical range between the
LocataLite A and the rover, trop
τ
is the error due to
tropospheric propagation effect and A T∂ is the clock
drift of the transmitter. The tropospheric model used is
the RTCM LASS model for radio communication. trop
τ
is proportional to a gradient of temperature according to
(Barnes, 2006b). However, for most of the LocataLite
networks the height difference between the rover and the
transmitters is in the order of magnitude of 50 meters
maximum. Thus, the tropospheric error may be negligible
in those measurements. j
A
N is the integer ambiguity an d
j
φ
ε is the propagation error on the phase measurement
(L1) (i.e. multipath, scattering). The static tests show that
the precision achieved with the 10 MHz spread spectrum
code is roughly 3m. This result is due to multipath
resolution for this kind of code (Hoffmann-Wellenhof et
al., 2001).
It is well known that the carrier-phase measurement is
much more precise than the pseudorange measurement.
Roughly, the carrier-phase error is proportional to 0.01
cycle or 0.001m (Hoffmann-Wellenhof et al., 2001), but
the disadvantage is that one more unknown variable (the
integer ambiguities) for each transmitter has to be
estimated in addition to the estimation of the receiver's
clock drift and the coordinates of the rover. At the time of
writing, the Locata technology uses only a single

Montillet: Achieving centimetre-level positioning accuracy in urban canyons with Locata technology 161
frequency to transmit the data, and this does not yet allow
On-The-Fly ambiguity resolution.
In a pre-processing step, the rover is initialized statically
on a precisely known point in order to calculate the
integer ambiguities. Thus, the integers remain constan t as
long as the carrier tracking loop maintains lock. Any
break in tracking, no matter how short, could change the
integer values. This happens if the rover and the
transmitter are not in a Line-of-Sight (LoS) or if the rover
enters in a deep fading zone (i.e. obstruction by trees,
strong scatterers).
2.4 Network Synchronization
The LocataNet currently in use is a Master/Slave
structure. All the Slave LocataLites are synchronised with
the reference PRN of the Master LocataLite, generally
PRN1. When designing their LocataNet, the user
manually decides which base-station is either a Master or
a Slave during the network setup. The synchronisation
process is called TimeLoc. If the Slave cannot be directly
synchronised with the Master, due to Non Line-Of-Sight
(NLOS) or a deep fading area, the synchronisation can be
done with another Slave by “cascading” synchronisation.
This can greatly simplify setup of networks in difficult
environments such as urban areas. Finally, the
transmitter’s clock offset has to be corrected. This
correction is called the network synchronisation.
Fig. 3 The differen t steps in the synchronization o f the LocataNet
(Barnes, 2006b)
Fig. 3 describes the TimeLoc process. Let us start with
the explanation of the pseudorange synchronisation: the
Master is named A, and B is the Slave LocataLite.
Beforehand, it is worth underlining that the TimeLoc
process for each LocataLite only involves the reference
PRN transmitter (Master), the top transmitter and receiver
from the Slave LocataLite. That means the two
transmitters of the same LocataLite share the same clock.
In the first step, the Master transmits a signal received by
the Slave. For the sake of clarity, the time delay (A T) is
due to the separation distance Tx-Rx (A R) and the clock
drift of the master (A t∂) (we neglect any troposphere
effects and random errors due to signal propagation).
Then, the Slave transmits its own signal at low power in
order to avoid the near-far effect phenomenon with the
signal coming from the Master and receives it (Fig. 3 -
Right). This time delay (B T) is correlated with the
Slave's transmitter clock offset. Mathematically the
problem can be viewed as:
cRtT AAA /+∂= (2)
cRtT BBB /+∂= (3)
The Slave subtracts the two time delays (A T-B T) and it
corrects this value by the geometrical ranges as it knows
the position of the transmitters from the memory card.
The reader may wonder how the coordinates of the
receiver are calculated. The author assumes the
transmitter and the receiver from the Slave LocataLite are
coupled, then the separation distance (B R) is equal to
zero. A R is calculated, replacing the coordinates of the
receiver by those of the Slave's transmitter.
Finally, using Direct Digital Synthesis (DDS) technology
(Barnes et al., 2003c), the Slave's transmitter adjusts its
local oscillator in order to have:
BAttt
∂
−
∂
=
∂
(4)
The Master and Slave are now synchronized. An
interesting example is to consider the special case when
the antennas (transmitter and receiver) are exactly on the
same vertical pole and the clock drift of the Master is
very small compared to the pseudorangeA
T. The
geometrical range can be calculated from the equations
(2) and (3) as:
(
)
(
)()( )
2222
AAAA cTZZYYXX =−+−+− (5)
(
)
(
)( )
=−+−+− 222 ZZYYXX BBB
()()
2
BB tTc∂− (6)
Where [AAA ZYX ,,], [BBB ZYX ,,] and [ZYX ,, ] are
the coordinates of the Master, the Slave's transmitter and
the Slave's receiver. c is the velocity of light. The
distance between the Slave's transmitter and Slave's
receiver is very small (A R>> B R). Finally, LocataLite B
removes the separation delay due to Master-Slave
separation. As it is assumed that the Slave’s antennas are
on the same pole, [X, Y] is equal to [BB YX ,]. Then the

162 Journal of Global Positioning Systems
clock synchronisation of the Slave is achieved by solving
the following equations:
B
tt ∂=∂ (7)
BAB cTZZ
c
t+−=∂ (
1
() ))()( 22
2BABAA YYXXcT −+−− (8)
The TimeLoc process of the carrier-phase code is
carried out after the pseudorange is synchronized. The
carrier-phase code synchronisation starts with the
calculus of the integer ambiguities for each LocataLite.
λ
AAAA cNcRtT ++∂= /
ˆ (9)
λ
BBBBcNcRtT ++∂= /
ˆ (10)
() ()
BABABAttTTcNN ∂−∂−−=− ˆˆ
λ
() cRRBA /−
−
(11)
In equations (9) and (10), A N and B N are the integer
ambiguities of the Master and Slave carrier-phase code
signal;
λ
is the wavelength of the signal. First, the
quantity (BANN −) is calculated replacing (A t
∂
-B t
∂
)
the quantity calculated using the pseudoranges (i.e. (4)).
As a matter of fact, the value found for the quantity
(BA NN−) is a float. Thus, in an iterative process the
clock of the Slave decreases this quantity and checks if it
can synchronize with the Master (using DSS). Finally, the
Master and the Slaves are synchronized at the sub-
nanosecond level. TimeLoc is not just one step in the
LocataNet configuration; it is a continuous process due to
the clock of the Master drifting over the time.
3 Experimental Setups & Results
Over the last five months, the authors have been making
several measurement campaigns with LocataLites around
the campus of the University of Nottingham (University
Park) in order to test and analyse the performance of the
Locata technology. Once the network is synchronized,
generally in several minutes, the rover takes the
pseudorange and carrier-phase measurements and the
navigation software triangulates a position - either by
post-processing, or in real time if required. This software
is called LINE (Locata Inline Navigation Engine).
Throughout this section, the results are only extracted
using the carrier-phase measurements, in order to show
the potential application of Locata technology to solve
the stringent positioning problems in severely obstructed
areas for the VISTA project.
3.1 The Navigation Software (LINE)
The LINE navigation software follows a three-stage
process: Measurements Correction, Select LocataLites
and Navigation Algorithm. In the first step, the carrier-
phase measurements and the pseudoranges have to be
corrected to take out some biases (roll over) due to
technical features. The software also detects if there is
any cycle slips in the carrier-phase measurements. When
detected, LINE repairs it. We set up the threshold for the
cycle slip detection at 0.333 cycles (~4 cm). The function
Select LocataLites is included in the software to detect
which LocataLites are valid for the trilateration of the
rover’s position. The main parameter is the SNR
recorded by the rover for each LocataLite. If the software
detects that the SNR value is under a specified threshold
(recorded in the memory card of the rover), it then
discards the LocataLite in the position computation for
this measurement epoch. In the final stage the Navigation
Algorithm uses a Least-squares algorithm to triangulate
the rover’s position (Strang and Borre, 1997). The
software may triangulate the position of the rover in 2D if
the 3D position diverges.
3.2 Static Measurements
Static tests show the accuracy of the LocataLites and the
evolution of any error epoch-by-epoch. The different
environments sum up the various applications where
Locata technology may be used in the future. In this part,
the results from four trials carried out at different
locations are presented: the downs on April 5 2007, the
parking area of the IESSG on 22 February 2007, and a
courtyard close to the George Green Library (GGL) on
March 28 2007 and April 26 2007. The first two places
represent light multipath environments, whereas the GGL
courtyard is surrounded by multiple scatterers such as
trees and buildings. The trial carried out at the IESSG car
park was done at daybreak when only a few cars moved
in or around the network. Fig. 4 is the orthophoto of the
main campus of the University of Nottingham.
The same reference system as GPS (WGS84) is used to
determine the coordinates of the rover and LocataLites
but these coordinates are further converted into Easting
Northing and Up coordinates in a local coordinate
system. For some reason, the navigation software was
switching the rover position computation from 3D to the
horizontal 2D coordinates (North and East). One potential
reason may be that the Vertical Dilution of Precision
(VDOP) value is too high (Massat & Rudnick, 1990).

Montillet: Achieving centimetre-level positioning accuracy in urban canyons with Locata technology 163
The results displayed in Table 1 are in millimetres and
they are obtained using only carrier-phase measurements.
Fig. 4 Overview of the U n i v e rsity of Nottingham with the different
places where the experiments took place
For each scenario, the results are averaged on 4000
epochs, and between 5 and 8 LocataLites were set up for
the experiments. Mean (Mu) and standard deviation (Std)
are calculated for the 2 Dimensions.
(mm) RMSE East North
Mu 45 6.7 44 Downs
(1.57) Std 9 6.5 9.9
Mu 38.1 9.2 13.5 GGL1
(0.63) Std 10 8.5 8.6
Mu 40 21 33.8 GGL2
(0.59) Std 11 6.4 10
Mu 10 4.6 8.5 Car
Park
(0.635) Std 4.7 2.8 1
Table 1: Position Accuracy - Static Scenarios
The minimum accuracy is around 1cm (Car Park) and the
worst is 4cm (GGL2). The results are down to the
millimetre level averaging on one coordinate
(i.e. mmMuDOWNS
EAST 7.6=).
Fig. 5 Time Series of the Rover' s Po sition [GGL2]
Fig. 5 shows the evolution of the total error (Root Mean
Square Error) and the error for the Easting and Northing
coordinates. It is clear that the static error remains stable
over the epochs (there are no biases).
3.3 Kinematic Tests
A kinematic test was also carried out on the university
campus. The results extracted are based on two different
scenarios: the first one has already been explained under
the name GGL1 and it has been performed in a courtyard
at the University of Nottingham, whereas the second
scenario took place in a courtyard at the University of
New South Wales in Australia (UNSW).
Fig. 6 The courtyard at the UNSW
Fig. 6 is the view from the top of the courtyard at the
UNSW surrounded by buildings, and 8 LocataLites were
used to perform the experiments.

164 Journal of Global Positioning Systems
GGL1 (mm) RMSE East North
Mu 68.5 25.4 1.36
TS Std 5.35 2.3 1
Mu 147.8 24.2 63
A1 Std 3.4 7.47 14
Mu 86 24.5 82.4
A2 Std 4.1 3.79 4.05
Mu 29.2 22.4 17.5
TS Std 8.34 6.09 8.87
Table 2: Kinematic test Results - GGL1
In the scenario GGL1, the kinematic test starts at the
point TS to initialize the LOCATA technology and then
the rover moves to A1 and further away to A2 then
finally it comes back to TS. The distance between TS and
A1 is 10 meters and TS-A2 is approximately 20 meters.
In two dimensions, the results are around 2 centimetres of
accuracy at the initial point. Moving further away the
accuracy degrades, reaching 14.7 centimetres on average
at A1 and approximately 8.6 centimetres at A2. Th e error
is around 3 cm when the rover comes back to the initial
point.
This experiment shows that there is no apparent
relationship between the distance and the error
propagation, because once the integer ambiguities have
been calculated for each signal transmitted by the
LocataLites, it remains constant and no update is
performed during the measurements. If a loss of lock
occurs during the test, the rover stops triangulating its’
position and the user needs to restart the initialization at
the starting point. The analysis of the estimated position
time series concludes that the fading environment around
the points is the main error source.
UNSW (mm) RMSE East North
Mu 8.6 8 2.7
TS Std 1.86 1.6 1.7
Mu 86.8 12.8 85
B5 Std 9.5 9.36 9.4
Mu 36.9 14 16.1
B6 Std 10.7 13.2 10.1
Mu 24 17 16.1
TS Std 10.4 9.8 5.76
Table 3: Kinematic test Results- UNSW
In the second scenario (UNSW), the rover is initialized
on the TS point and moves to B5, B6 and finally back to
the first position. The distance between TS and B6 is
5 m, and TS-B5 is 11.5 m.
Fig. 7 Time Series of the Rover's Po sition (UNSW)
In two dimensions (Northing and Easting), the accuracy
is comparable to the first scenario with a mean error of
1.5 centimetres. Fig. 7 shows the time series of the RMS
error at the point B6, and then whe n we co me b a ck to TS.
Although the HDOP is around 0.56, the accuracy
degrades when moving from the initial point: the
maximum error is 8.5 centimetres on the North
coordinate at B5. This result confirms that both the
network configuration and the fading environment are the
two main parameters in order to explain the accuracy for
each geom e t ri c al point.
Finally, an analysis of the vertical component is not
included in this paper because this aspect of the study is
still being researched.
4 Future Work
This work gives an insight into the Locata technology.
Several trials have been co nducted around the University
of Nottingham to test this new technology, with a
comparison given for different environments (open and
built-up areas). An analysis of the results shows that for
both kinematic and static tests, the 2 dimensional errors
(East and North coordinates) are of the order of
magnitude of the centimetre level of accuracy. The
maximum error is observed in kinematic tests with a
value close to 15 cm, and the minimum is a few
millimetres in static tests. This kind of accuracy
demonstrates that Locata technology is a suitable
candidate, and a promising technology, to precisely locate
buried pipes and assets in urban canyons.
Researchers are still investigating the error on the third
dimension. To do so, a LocataNet has just been installed
(July 2007) which consists of LocataLite stations on the

Montillet: Achieving centimetre-level positioning accuracy in urban canyons with Locata technology 165
roofs of the buildings and ground points surrounding the
GGL courtyard at the University of Nottingham. This
setup should significantly decrease the VDOP value of
the network. Researchers are also investigating Wi-Fi
interference, an unexpected phenomenon which was
observed during some of the previous measurement trials.
Acknowledgements
The authors would like to thank: Locata Corp for their
advice about how to setup the network and especially Mr
Nunzio Gambale and Dr. Jimmy LaMance from Locata
Corp; Mr Noni Politi (from the UNSW), Mr Huib De
Ligt and Mr Richard Barnes (from the IESSG) for their
help during the experiments. This work is supported by
the Visualizing Integrated Information on Buried Assets
to Reduce Streetworks (VISTA), project funded by the
UK DTI.
References
Roberts G., Meng X., Montillet ,J-P. & Taha A. (2006) The
Location and Positioning of Buried Pipes and Cables in
Built Up Areas, Proceedings of XXIII Fig Congress:
Shaping the Change, October, 1-9.
Parker, J. (2006) Minimising Street Works Disruption: Buried
Asset Data Collection and Exchange Field Trials,
UKWIR Report 06/WM/12/9.
Boukhelifa, N. & Duke D. J. (2007) The Uncertain Reality of
Underground Assets, Joint Workshop "Visualization and
Exploration of Geospatial Data", ISPRS, ICA, DGFK. June
27-29.
Montillet J-P, Taha A., Meng X. & Roberts G. W. (2007)
Buried Assets: Testing GPS and GSM in Urban Canyons,
GPSWorld Magazine, March, 39-43.
Barnes J., Rizos C., Kanli M. & Pahwa A. (2006) Locata: a
New Positioning Technology for Classically Difficult
GNSS Environments, International Global Navigation
Satellite Systems Society IGNSS Symposium 2006, July.
Prasad R. (1998) Universal Wireless Communications, Artech
House Publisher, 1st Edition.
Cobb H. Stewart (1997) GPS Pseudolites: Theory, Design and
Applications, PhD thesis, Stanford University.
Cosser Emily (2004) Bridge Deformation Monitoring With
Single Frequency GPS Augmented by Pseudolites, PhD
thesis, University of Nottingham, Institute of Engineering
Surveying and Space Geodesy.
Misra P. & Enge P. (2001) Global Positioning System: Signals,
Measurements, and Performance, Ganga-Jamuna Press,
1st Edition.
Kanli M. O. (2005) Limitations of Pseudolites Systems Using
Off-The-Shelf GPS Receivers, Journal of Global
Positioning Systems, 3(1-2), 154-166.
Van Dierendonck, A.J. (1997) Satellite Radio Navigation, in
Avionic Navigation Systems, Myron Kayton nad Wlater R.
Fried (Eds.), John Wiley \& Sons, 2nd Edition.
Barnes J., Rizos C., Wang J., Small D., Voigt G & Gambale N.
(2003) Locata: A New Positioning Technology for High
Precision Indoor and Outdoor Positioning, Proceedings
2003 International Symposium on GPS\ GNSS, 9-18.
Yun D. & Kee C. (2002) Centimeter accuracy stand-alone
indoor navigation system by synchronized Pseudolite
constellation, Proceedings of the 15th International
Technical Meeting of the Satellite Division of the Institute
of Navigation ION GPS.
Novariant (2005) Novariant terralite XPS deployment at
Phelps Dodge Morenci mine, technical report.
Montillet J-P. (2007) Achieving the centimetre level of
Accuracy using Locata Technology, Proceedings of the
ION GNSS Symposium, fort-worth, Texas.
Proakis J.G. (2000) Digital Communications, Mc Graw-Hill,
New York, NY, 4th Edition.
Barnes J., Rizos C., Wang J., Small D., Voigt G., & Gambale N.
(2003b) Locata: the Positioning Technology of the
Future? Proceedings of the 6th International Symposium
on Satellite Navigation Technology Including Mobile
Positioning and Location Services, July.
Barnes J. (2006b) Ground-breaking Ground-based positioning,
Locata Training Course, technical report.
Hoffmann-Wellenhoff B., Lichtenegger H. & Collins J. (2001)
GPS: Theory and Practice, Springer-Verlag/Wien, 5th
Edition.
Barnes J., Wang J., Rizos C., Small D., Voight G. & Gambale
N. (2003c) Guidelites: Intelligent synchronised
pseudolites for cm-level stand-alone positioning,
Proceedings 11th Int. Assoc. of Institutes of Navigation
(IAIN) World Congress, October, 21-24.
Strang G. & and Borre K. (1997) Linear Algebra, Geodesy, and
GPS, Wellesley Cambridge Press, Wellesley, MA.
Massat P. & Rudnick K. (1990) Geometric formulas for
dilution of precision calculations, Journal of the Institute
of Navigation, 37(4), 379-391.