Documentos de Académico
Documentos de Profesional
Documentos de Cultura
th
The 6 International Symposium on
Satellite Navigation Technology Including
Mobile Positioning & Location Services
Melbourne, Australia
2225 July 2003
ABSTRACT
Locata Corporation has invented a new positioning technology called
Locata, which is designed to overcome severe limitations in other
positioning systems currently available. Part of the Locata technology
consists of a time-synchronised pseudolite transceiver called a LocataLite.
A network of LocataLites forms a LocataNet, which transmits GPS-like
signals that allow single-point positioning using carrier-phase measurements
for a mobile device (a Locata). The SNAP group at UNSW has assisted in
the development of a Locata and testing of the new technology. In this
paper, the prototype Locata technology is described, and the results of a
performance test experiment are presented.
The test experiment
demonstrates proof-of-concept for the Locata technology and shows that
carrier point positioning (without radio modem data-links) is possible with
sub-centimetre precision.
KEYWORDS: High-precision, Kinematic-positioning, Time-synchronised
1. INTRODUCTION
GPS is used for precise centimetre-level positioning in a variety of outdoor applications,
where a relative carrier-phase differential technique (using a base station) is almost
exclusively used. Differential operation, most commonly using the double-differenced
observable, is necessary to reduce orbit errors, spatially correlated errors due to the
atmosphere, and to eliminate both receiver and satellite clock biases. For real-time
positioning (i.e., RTK), the roving GPS receiver must receive base station data via a radio
PAPER 49 (peer-reviewed)
modem (or other wireless link) from a base station. This, combined with the fact that RTK
GPS only works well with a relatively unobstructed and geometrically good satellite
constellation, and the operating range of a rover receiver is typically limited to less than 10km
due to ionospheric effects, is a significant limitation of the RTK-GPS technology.
In situations where GPS satellite geometry is poor or the signal availability is limited, groundbased transmitters of GPS-like signals (called pseudolites) can be used to augment GPS.
They therefore have the potential to be used for both outdoor and indoor positioning. With
enough pseudolites it is theoretically possible to replace GPS entirely, though in practice this
has been difficult to achieve. Typically pseudolites use cheap crystal oscillators and operate
independently (in the so-called unsynchronised mode). In this case, double differencing
must be used to eliminate the pseudolite and receiver clock biases.
The SNAP group has been conducting pseudolite research for the past three years and
experimenting with them in the unsynchronised mode for a variety of applications (Barnes et
al. 2002a, Barnes et al. 2002b, Wang 2002, Wang et al. 2001, Dai et al. 2001). Therefore,
real-time centimetre-level positioning with unsynchronised pseudolites can only be achieved
with a base station that provides data to a rover unit via a radio modem (as with standard
RTK-GPS). If pseudolites can be synchronised, stand-alone positioning can be achieved
without base station data (and without the radio modem data link). Until now attempts to
synchronise pseudolites have resulted in position solutions that are up to six times worse in
comparison to an unsynchronised approach using double-differencing (Yun and Kee, 2002).
Locata Corporation has invented a new positioning technology (Locata), that consists of a
network (LocataNet) of time-synchronised pseudolite transceivers (LocataLites). In the
following sections, the Locata technology is described, and real-time stand-alone (without
base station data) high precision (sub-cm) point positioning is demonstrated.
2.1 LocataLite
The LocataLite can be described as an intelligent pseudolite transceiver. The transmitter
prototype hardware used is such that the intelligence of the unit is in its software. This is an
extremely flexible approach and allows major design changes without requiring completely
new hardware. The receiver part of the prototype is based on an existing GPS receiver
chipset, which is described in section 2.3. The receiver chipset and the transmitter share the
same clock, which is a cheap temperature-compensated crystal oscillator (TCXO). The
transmitter part of the prototype generates C/A code pseudorange and carrier-phase signals at
the GPS L1 frequency. The signal is generated digitally (unlike most existing pseudolites,
PAPER 49 (peer-reviewed)
which use analogue techniques) and can be operated in a pulsing mode with different duty
cycles, power output, and any PRN codes can be generated. Pulsing is commonly used with
pseudolite signals (instead of a continuous transmission, like GPS), to reduce interference and
increase the working range (the near-far problem). The duty cycle refers to the percentage of
time the pseudolite is transmitting when pulsing. Commercially available GPS patch
antennas are used for the receiver and transmitter, in addition to a custom built wave
antenna for one of the LocataLite transmitters. The prototype LocataLite and antennas are
shown in Figure 1.
2.2 TimeLoc
In order for a mobile receiver to carry out carrier-phase point positioning without the need for
base station data, the LocataLite devices must be time-synchronised. The level of
synchronisation required is extremely high, considering a one nanosecond error in time
equates to an error of approximately thirty centimetres (due to the speed of light). The timesynchronisation procedure of one or more LocataLite devices is a key innovation of the
Locata technology and is know as TimeLoc. The TimeLoc procedure to synchronise one
LocataLite (B) to another LocataLite (A) can be broken down into the following steps:
1. LocataLite A transmits a C/A code and carrier signal on a particular PRN code.
2. The receiver section of LocataLite B acquires, tracks and measures the signal (C/A code
and carrier-phase measurements) generated by LocataLite A.
3. LocataLite B generates its own C/A code and carrier signal on a different PRN code to A.
4. LocataLite B calculates the difference between the code and carrier of the received signal
and its own locally generated signal. Ignoring propagation errors, the differences between the
two signals are due to the difference in the clocks between the two devices, and the geometric
separation between them.
5. LocataLite B adjusts its local oscillator using Direct Digital Synthesis (DDS) technology to
bring the code and carrier differences between itself and LocataLite A to zero. The code and
carrier differences between LocataLite A and B are continually monitored so that they remain
zero. In other words, the local oscillator of B follows precisely that of A.
PAPER 49 (peer-reviewed)
6. The final stage is to correct for the geometrical offset between LocataLite A and B, using
the known coordinates of the LocataLites, and after this TimeLoc is achieved.
Importantly, the above procedure does not require expensive atomic clocks, and there is in
theory no limit to the number of LocataLites that can be synchronised together using
TimeLoc.
2.3 A Locata
To speed up the development of a prototype system it was decided to use existing GPS
hardware for the receiver section in the LocataLite and the Locata (the mobile positioning
device). The SNAP Group at UNSW has assisted in the development of the Locata through
Mitels (now Zarlink) GPS Architect development system (Zarlink, 1999). The development
system uses the Mitel GP2000 chipset comprised of the GP2015 RF front end and GP2021
12-channel correlator together with the P60ARM-B microprocessor (Ibid). Importantly the
system includes GPS firmware C source code that can be modified, compiled and uploaded to
the GPS receiver. However, the GPS Architect hardware is designed as an indoor laboratory
development tool and not suited to outdoor use.
Instead of designing and building GPS receiver hardware (using the GP2000 chipset), suitable
for outdoor use, a different approach was taken. This was to modify a Canadian Marconi
Corp (CMC) Allstar GPS receiver, which uses the Mitel GP2000 chipset, so that it would
operate in exactly the same way as the GPS Architect hardware. The original GPS Architect
firmware source code has been extensively modified and improved, by the Locata
Corporation and the SNAP group. The modifications have been in signal acquisition, the
tracking loops and the navigation algorithm. The prototype Locata hardware and antenna (a
commercially available patch antenna) are shown in Figure 2.
GPS. The basic GPS L1 carrier-phase observation equation between receiver A and satellite j
in metres can be written as:
Aj = Aj + trop + c TA c T j ion
c
N Aj +
f L1
(1)
Aj = Aj + trop + c TA
j
c
N Aj +
f L1
(2)
where the terms are the same as for GPS, except they refer to LocataLites instead of satellites.
In equation 2 there is no clock error due to the LocataLites since they are time-synchronised
to each other (see Section 2.2), and because the devices are ground based there is no
ionospheric correction term. The tropospheric correction will depend on the separation
between the Locata and the LocataLite, the elevation angle to the LocataLite, and the
atmospheric conditions (pressure, temperature, humidity and pressure) along the line-of-sight
signal path.
The term that poses the most difficulty in the above equation is the unknown number of
carrier wavelengths between the Locata and the LocataLite when TimeLoc is achieved. In the
PAPER 49 (peer-reviewed)
prototype system, the ambiguity term and the initial receiver clock error are determined
through a static initialisation at a known point. Assuming that the tropospheric effects are
modelled or negligible due to relatively short distances between the Locata and LocataLite,
the initial bias (clock error and ambiguity) in metres can be written as:
BAj = c TA
j
c
N Aj +
f L1
(3)
BAj = Aj Aj
(4)
Aj = Aj + BAj + dTA +
(5)
Aj = ( X A X j ) 2 + (YA Y j ) 2 + ( Z A Z j ) 2
(6)
and
where dTA is the change in the receiver clock error from the static initialisation epoch, and
this together with the Locata coordinates X A , YA , Z A give four unknowns; which can be
solved with a minimum of four LocataLite carrier-phase measurements and least squares
estimation. The least squares estimation procedure is similar to that for standard GPS single
point positioning (SPP), except that the very precise carrier-phase measurement is used. After
the carrier-phase bias is determined through static initialisation the Locata is free to navigate
kinematically. The positioning algorithm is embedded in the GPS firmware of the Locata to
allow for real-time positioning. It should also be stressed that each positioning epoch is
independent and no smoothing or filtering is carried out in the prototype system.
PAPER 49 (peer-reviewed)
PRN
used
Transmit/Receive
Antennas
Master
North
East
South
West
32
12
14
21
29
wave/NA
Patch/Patch
Patch/Patch
Patch/Patch
Patch/Patch
Elevation angle
from rover pole
(Degrees)
24.5
3.4
9.1
5.2
-4.0
Distance from
rover pole (m)
5.5
36.9
103.8
29.6
108.1
SNR
mean/stdev
(unit)
20.1/0.019
20.1/0.051
21.2/0.101
19.8/0.041
18.8/0.080
Single
difference
stdev (mm)
Reference
9.6
8.1
6.6
7.0
Table 1. LocataLite trial details: elevation angle and distance from rover pole, SNR and single
difference statistics.
PAPER 49 (peer-reviewed)
80
60
40
north
North (m)
west
20
master
rover
0
20
south
east
40
60
80
100
80
60
40
20
20
40
60
80
100
East (m)
PAPER 49 (peer-reviewed)
The signal-to-noise ratio (SNR) values of the five LocataLite units, recorded by the Locata,
during the static test are given in Figure 6. Also, the mean and standard deviation of the SNR
time series are given in Table 1. Because the LocataLite units and the Locata are stationary it
is expected that the SNR values should be random with a constant mean, unlike GPS SNR
values which typically increase as the satellite elevation angle increases. Overall, the signal
strength from all the LocataLites are good and similar, with mean values ranging from 18.8 to
21.2 dB. The variation of the SNRs is relatively small, with the absolute variation less than
0.5 dB for all LocataLites. This indicates that the signal power from the LocataLites is stable
and that the Locata is able to track the signal without difficulty.
It is interesting to note that there does appear to be some correlation between standard
deviation of SNR values and the distance between the LocataLites and the Locata (Figure 6).
The SNR values from the Master, only 5.5 metres from the Locata have the smallest standard
PAPER 49 (peer-reviewed)
deviation (least variation) of 0.02 dB. The SNRs of the East (PRN 14) and West (PRN 29)
LocataLites have the largest variation with a standard deviation of approximately 0.1 dB.
These devices are the greatest distance from the Locata (approximately 100 m).
SNR 32
22
21.5
21
SNR
20.5
20
19.5
19
18.5
18
0
200
400
600
800
1000
Epoch (s)
1200
1400
stdev 0.019
1600
1800
2000
mean 20.06
22
21.5
21.5
21
21
20.5
20.5
20
20
19.5
19.5
19
19
18.5
18.5
18
0
200
400
600
800
Epoch (s)
1000
1200
1400
stdev 0.051
SNR 21
1600
1800
18
0
2000
400
600
800
Epoch (s)
22
22
21.5
21.5
21
21
20.5
20.5
20
19.5
1000
1200
1400
stdev 0.101
SNR 29
1600
1800
2000
mean 21.21
20
19.5
19
19
18.5
18.5
18
0
200
mean 20.06
SNR
SNR
SNR 14
22
SNR
SNR
SNR 12
200
400
600
800
Epoch (s)
1000
1200
1400
stdev 0.041
1600
1800
mean 19.81
2000
18
0
200
400
600
800
Epoch (s)
1000
1200
1400
stdev 0.080
1600
1800
2000
mean 18.80
A good way to assess how well the LocataLite units are time-synchronised is to compute
single-difference measurements between the LocataLites. This will eliminate the Locata
clock error, and show any errors between the LocataLite clocks. Using the logged
measurement data, single difference observations were computed between the Master and all
other LocataLites. The ambiguities of the single differences were resolved using the known
coordinates of the LocataLites and the rover pole.
Figure 7 shows the four single differences between the Master and the other LocataLites.
Most importantly, visually all the single difference time series on average fit a horizontal line
and do not have any long-term drifts during the thirty-five minute test. The overall standard
deviations of the single difference time series are all less than 1 cm (see Table 1), and in terms
of how well the LocataLite clocks achieve TimeLoc this equates to better than 33 picoseconds. However, visually the time series do not appear entirely random and the cause of the
fluctuations requires further investigation. Also interesting to note is that the standard
deviations of the time series do not appear to be correlated with the standard deviations of the
SNRs, and is worthy of further investigation.
PAPER 49 (peer-reviewed)
10
0.04
0.04
0.03
0.03
0.02
0.02
0.01
0.01
Metres
Metres
0
0.01
0
0.01
0.02
0.02
0.03
0.03
0.04
0.04
0.05
0
0.05
200
400
600
800
1000
1200
1400
1600
1800
2000
200
400
600
0.05
0.05
0.04
0.04
0.03
0.03
0.02
0.02
0.01
0.01
0
0.01
1000
1200
1400
1600
1800
2000
1800
2000
0
0.01
0.02
0.02
0.03
0.03
0.04
0.04
0.05
0
800
Epoch (s)
stdev 8.1mm
L1 single difference 2932
Metres
Metres
Epoch (s)
stdev 9.6mm
L1 single difference 2132
0.05
200
400
600
800
1000
1200
Epoch (s)
1400
1600
1800
2000
200
400
600
stdev 6.6mm
800
1000
1200
Epoch (s)
1400
1600
stdev 7.0mm
To assess the accuracy of the real-time positioning results, the known (sub-cm) coordinate of
the rover pole was used to compute the positioning error for each epoch. Figure 8 shows the
East and North errors for the real-time positions of the Locata. The mean error of the both
time series is less than 2 mm, with the standard deviations and root-mean-square values less
than 6 mm. Clearly sub-centimetre positioning precision has been achieved with 93% of the
East and North errors less than 1 cm. As expected from the analysis of the single difference
residuals, there are no long-term drifts in the time series (Figure 7). Also as expected,
visually the time series do not appear entirely random due to the patterns in the single
difference residuals as noted previously. The above results demonstrate that CPP with subcentimetre precision is clearly achieved with the Locata technology.
North
0.05
0.04
0.04
0.03
0.03
0.02
0.02
Error (metres)
Error (metres)
East
0.05
0.01
0
0.01
0.01
0
0.01
0.02
0.02
0.03
0.03
0.04
0.04
0.05
0.05
0
200
400
600
Epoch (s)
800
1000
1200
stdev 4.5 mm
1400
1600
mean 1.8 mm
1800
2000
rms 4.8 mm
200
400
600
Epoch (s)
800
1000
1200
stdev 5.5 mm
1400
1600
mean 0.2 mm
1800
2000
rms 5.5 mm
11
box mounted on the turntable houses the Locata, battery and laptop computer for data
logging. The antenna was mounted on a bar attached to the top of the box, giving a seventycentimetre circle radius.
During the test, lasting sixteen minutes, the Locata antenna rotated at an approximate linear
velocity of 2.4 m/s, and an angular velocity of 3.4 rad/s. Both real-time positions and raw
measurement data were recorded using a laptop via a serial interface. Before the antenna
started to rotate, the carrier biases were determined through initialisation at the rover pole
with known coordinates.
To assess how close the Locata coordinates lie on a circle, a least squares procedure was used
to estimate the radius and centre of the circle using the Easting and Northing data. The
computed radius differs to that measured (using a steel tape measure) by less than 3 mm, and
Figure 10 shows the horizontal position of the Locata. Visually, the plot depicts a circle and
suggests that North-South precision is slightly better than East-West, and is partly due to the
fact that the dilution of precision North-South is approximately 28% better than East-West.
Figure 10 also shows the least squares residuals of the circle estimation using the position
data. The overall standard deviation of the residuals is 16 mm, while 82% of values are less
than 20 mm from the best-fit circle. There are occasional residual values as large as 60 mm
in the time series, which could be due to multipath from people walking around the test area
during the test or parked vehicles. Also, since the circle repeating test equipment is not
precisely calibrated, levelled or ridged, then this will impact on the accuracy of the
positioning results. Despite the errors induced by possible multipath and the test equipment,
centimetre-level kinematic positioning can clearly be achieved using the Locata technology.
PAPER 49 (peer-reviewed)
12
Circle centre
0.1890
0.6659
radius
0.6971
0.1
0.8
0.08
0.6
0.06
0.4
0.04
0.02
Residual (m)
Northing (m)
0.2
0.2
0.02
0.4
0.04
0.6
0.06
0.8
0.08
0.5
0
Easting (m)
0.5
0.1
100
200
300
400
Min 58.2 mm
500
Epoch (s)
600
700
Max 66.8 mm
800
900
Figure 10. Horizontal position plot (left) and least square residuals (right) of circle repeating
kinematic test.
4. CONCLUDING REMARKS
Is Locata the positioning technology of the future? In this paper it has been shown that the
prototype Locata technology clearly demonstrates the proof-of-concept of a timesynchronised network for positioning. At an outdoor test network of LocataLites
(LocataNet), approximately 200x60 metres, a mobile unit (a Locata) performed static carrierphase point positioning with sub-centimetre precision. Also, in a kinematic test centimetre
level precision was obtained despite errors induced by possible multipath and the test
equipment used. This level of precision is remarkable for a prototype system, and is as good
as (if not better than) GPS RTK using a base station, radio modem and double differencing.
A real-time positioning technology that can operate indoors and outside anywhere in the
world, with sub-cm accuracy, and low cost is the ultimate goal for many researchers. Locata
Corp, with the assistance of SNAP (in software development and testing), is continuing to
develop the Locata technology to achieve this goal.
REFERENCES
Barnes J, Rizos C, Wang J, Nunan T, Reid C (2002a) The development of a GPS/Pseudolite
positioning system for vehicle tracking at BHP Steel, Port Kembla Steelworks, Proceedings of the 15th
International Technical Meeting of the Satellite Division of The Institute of Navigation ION GPS
2002, Portland, Oregon, 1779-1789
Barnes J, Wang J, Rizos C, Tsujii T (2002b) The performance of a pseudolite-based positioning
system for deformation monitoring, Proceedings of 2nd Symp. on Geodesy for Geotechnical &
Structural Applications, Berlin, Germany, 326-327
Dai L, Rizos C, Wang J (2001) The role of pseudosatellite signals in precise GPS-based positioning,
Journal of Geospatial Engineering, 3(1), 33-44
Leick A (1995) GPS satellite surveying (second edition), John Wiley & Sons, Inc., 560pp
PAPER 49 (peer-reviewed)
13
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 2002, Portland, Oregon, 213-225
Wang J (2002) Applications of pseudolites in geodetic positioning: Progress and problems. Journal of
Global Positioning Systems, 1(1), 48-56
Wang J, Tsujii T, Rizos C, Dai L, Moore M (2001) GPS and pseudo-satellites integration for precise
positioning. Geomatics Research Australasia, 74, 103-117
Zarlink (1999) GP2000 GPS receiver hardware design application note. Zarlink semiconductor, 54pp
PAPER 49 (peer-reviewed)
14