scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Performance Analysis of Using the Next generation Australian SBAS with Precise Point Positioning Capability For Intelligent Transport Systems

09 Apr 2019-pp 1-4
TL;DR: The SBAS-based PPP solutions have shown to give the best positioning precision and accuracy among all tested solution types, with sub-decimeter level accuracy, provided that enough convergence time is available.
Abstract: In 2018, a next-generation Satellite-Based Augmentation System (SBAS) test-bed was launched in Australia/New-Zealand in preparation for building an operational system. This new generation SBAS includes Ll legacy SBAS, new dual-frequency multi-constellation (DFMC) SBAS, and orbit and clock corrections for precise point positioning (PPP) using GPS and Galileo. In this paper, the next generation SBAS and its models are first presented, and the benefits of using its new components are discussed. Test results for lane identification applications in Intelligent Transport Systems (ITS) are presented and analyzed. Kinematic tests were performed in different ITS environments. These are characterized by different levels of sky-visibility and multipath, including clear sky, suburban, low-density urban, and high-density urban environments. Performance analysis show that results vary widely depending on the operational conditions but all SBAS solutions have better positioning accuracy compared with the standalone solutions that are currently used in transport applications. The DFMC SBAS slightly outperformed the Ll SBAS, with accuracy at sub-meter, and it has advantages during periods of fluctuations of the ionosphere with an extended coverage area. As expected, the SBAS-based PPP solutions have shown to give the best positioning precision and accuracy among all tested solution types, with sub-decimeter level accuracy, provided that enough convergence time is available. The paper concluded by giving remarks on the use of this new technology for ITS.

Summary (2 min read)

I. INTRODUCTION (HEADING 1)

  • The Satellite-Based Augmentation System (SBAS) are used in different regions to augment Global Navigation Satellite System observations to improve positioning and navigation of single point positioning (SPP).
  • While the SBAS L1 service transmits range, orbit, clock and ionospheric delay corrections, applicable to the service area covering the complete Australian and New Zealand territories, the SBAS L5 DFMC provides GPS and Galileo dual-frequency augmentation that can be used anywhere across the whole footprint of the GEO satellite.
  • The AUS/NZ SBAS additionally broadcast precise satellite orbits and clock corrections to support real-time float-ambiguity Precise Point Positioning (PPP) service that can deliver 5-20cm accuracy [4, 5, 6] .
  • The next section briefly overviews the AUS/NZ SBAS and discuss processing of data.

II. DESCRIPTION OF THE AUS/NZ SBAS

  • The second-generation AUS/NZ SBAS test-bed was developed by Geoscience Australia (GA) and Land Information New Zealand (LINZ), and is administrated by the Cooperative Research Centre for Spatial Information (now known as FrontierSI).
  • The SBAS infrastructure consists of the space segment, the ground segment, and the support segment in addition to the user segment (see Fig 1) .
  • The Test-bed signal contains corrections and DFREI bounds applicable to GPS L1/L2 + GAL E1/E5a ionosphere-free combinations.
  • Iii. Real-time PPP (RT-PPP) through SBAS L1 and SBAS L5: The AUS/NZ solution uses the spare bits present in the SBAS message to provide additional information for PPP with a higher resolution.
  • The ambiguity parameters and the ZTDs were modelled as random walk bias parameters.

IV. TEST DESCRIPTION

  • Testing of the second generation SBAS in this study was conducted with focus on two focus areas: i 'Heavy Vehicle Efficiency' to examine the SBAS-based positioning ability to improve transport network efficiency by identifying which lane the vehicle is travelling in, and ii.
  • The tests were carried out in July and August 2018 in Wollongong and Sydney, Australia.
  • The same raw code and phase observations used for SBAS-based positioning were utilized with data from Continuously Operating Reference Stations (CORS), serving as base stations, where the test vehicles were within a radius of eight kilometers from the base stations.
  •  Mean Error and RMSE were computed after all outliers were removed.
  • The gaps seen for some positioning modes refer to unavailable positions due to observing a limited number of satellites resulting from signal obstruction by structures or trees.

High-density Urban

  • Low-density Urban Suburban Fig. 5 . SBAS L1 results, suburban environment.
  • For real-time PPP, the figures show that the loss of visibility of some satellites led to a temporary/partial loss of convergence, and some time was needed to re-converge to the previous precision levels.
  • In addition, Figure 8 shows the mean absolute 2D horizontal error for the different positioning modes, which demonstrates average positioning errors and hence the system accuracy, regardless of the sign of the errors.
  • It can be seen that both L1 and DFMC SBAS gave much better positioning accuracy compared with SPP, and thus they can provide sub-meter horizontal positioning accuracy for road applications when sufficient number of satellites can be observed.
  • In the less dense urban environments, the availability for the SBAS modes was over 85% and over 90% for PPP.

C. Two vehicles results

  • In transport, one should not only consider the absolute positions of vehicles, but also their relative positions with respect to other surrounding vehicles sharing the road.
  • The cross-correlation between the errors of the two vehicles for different runs was from 0.35 to 0.55 for both L1 and DFMC SBAS and PPP.
  • The vehicle can be considered located in a certain lane when its position lies within an allowable range () from the centerline of the lane.
  • One should note that since the mean results are used, the test addresses a 50% probability.
  • PPP was acceptable when the number of visible satellites permitted convergence of the solution.

VII. CONCLUSION

  • Testing show that SBAS positioning in open sky environment and when observing satellites with good number and geometry can reach horizontal accuracy in the order of sub-meter.
  • The use of DFMC SBAS using GPS+Galileo slightly improves positioning accuracy compared to the legacy L1 SBAS.
  • The DFMC SBAS also provides extended coverage area and can provide better performance during periods of high fluctuations of the ionosphere.
  • This need for convergence makes the PPP solution less appealing for short journeys but it can be used for long trips, e.g. along highways.
  • These are common concerns in positioning using GNSS.

Did you find this useful? Give us your feedback

Content maybe subject to copyright    Report

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©2019 IEEE
Performance Analysis of Using the Next generation
Australian SBAS with Precise Point Positioning
Capability For Intelligent Transport Systems
Ahmed El-Mowafy
School of Earth and Planetary Sciences
Curtin University
Perth, Australia
a.el-mowafy@curtin.ed.au
Norman Cheung
HMI Technologies
Sydney, Australia
norman.cheung@hmitechnologies.com.au
Eldar Rubinov
FrontierSI,
Melbourne, Australia
erubinov@frontiersi.com.au
Abstract In 2018, a next-generation Satellite-Based
Augmentation System (SBAS) test-bed was launched in
Australia/New-Zealand in preparation for building an
operational system. This new generation SBAS includes L1
legacy SBAS, new dual-frequency multi-constellation (DFMC)
SBAS, and orbit and clock corrections for precise point
positioning (PPP) using GPS and Galileo. In this paper, the
next generation SBAS and its models are first presented, and
the benefits of using its new components are discussed. Test
results for lane identification applications in Intelligent
Transport Systems (ITS) are presented and analyzed.
Kinematic tests were performed in different ITS environments.
These are characterized by different levels of sky-visibility and
multipath, including clear sky, suburban, low-density urban,
and high-density urban environments. Performance analysis
show that results vary widely depending on the operational
conditions but all SBAS solutions have better positioning
accuracy compared with the standalone solutions that are
currently used in transport applications. The DFMC SBAS
slightly outperformed the L1 SBAS, with accuracy at sub-
meter, and it has advantages during periods of fluctuations of
the ionosphere with an extended coverage area. As expected,
the SBAS-based PPP solutions have shown to give the best
positioning precision and accuracy among all tested solution
types, with sub-decimeter level accuracy, provided that enough
convergence time is available. The paper concluded by giving
remarks on the use of this new technology for ITS.
KeywordsSBAS, PPP, GPS, Galileo, Intelligent Transport
Systems
I. INTRODUCTION (HEADING 1)
The Satellite-Based Augmentation System (SBAS) are
used in different regions to augment Global Navigation
Satellite System observations to improve positioning and
navigation of single point positioning (SPP). SBAS uses
geostationary (GEO) satellites to transmit orbit and clock
corrections, regional ionospheric corrections, ranging and
integrity information. Currently operational SBAS includes
the USA Wide Area Augmentation System (WAAS), the
European Geostationary Navigation Overlay Service
(EGNOS), the Japanese Multi-functional Satellite
Augmentation System (MSAS), the Russian System for
Differential Correction and Monitoring (SDCM) and the
Indian GPS-aided GEO Augmented Navigation (GAGAN).
These systems, except for MSAS, currently support only
single-frequency measurement users but in the future, such
as for EGNOS V3 will include DFMC service.
Since September 2017, Australia and New Zealand
(referred to from then on as AUS/NZ) commenced SBAS
test-bed over the Asia-Pacific and Australasia area [1]. This
SBAS comprises a second-generation SBAS service in a
pre-step of the building of a fully operational system within
a few years. The test-bed provides in addition to the L1
legacy SBAS signals that are transmitted over L1, the dual-
frequency multi-constellation (DFMC) SBAS corrections
being transmitted over L5. The former (i.e. SBS L1) meets
the RTCA/DO-229 standard and is open to current mass-
market receivers implementing SBAS positioning mode.
The second service; the new generation SBAS L5 DFMC
prototype signal, complies with the specification defined in
WG62 GAL GPS SBAS MOPS v0.6.1 [2]. The WG62 is a
draft standard that defines future DFMC services. Due to
bandwidth limitations, the AUS/NZ DFMC service only
supports GPS and Galileo satellites for which corrections
and integrity data are available [3]. While the SBAS L1
service transmits range, orbit, clock and ionospheric delay
corrections, applicable to the service area covering the
complete Australian and New Zealand territories, the SBAS
L5 DFMC provides GPS and Galileo dual-frequency
augmentation that can be used anywhere across the whole
footprint of the GEO satellite. For the AUS/NZ SBAS test-
bed, an Inmarsat 4F1 GEO satellite is used, which is located
at longitude of 143.4 degrees East [1].
The AUS/NZ SBAS additionally broadcast precise
satellite orbits and clock corrections to support real-time
float-ambiguity Precise Point Positioning (PPP) service that
can deliver 5-20cm accuracy [4, 5, 6]. The corrections are
transmitted over L1 frequency for GPS only and L5 for
GPS+Galileo. This service allows for superior availability
of PPP compared with traditional PPP. It enables PPP
anywhere in the footprint of the SBAS GEO satellite, i.e.
most Australia-Asia pacific region, whereas traditional PPP
depends on the Internet to receive the corrections, and thus
is limited to areas where the Internet is available. Noting the
vast size of low populated areas in Australia and the
surrounding oceans, such service of providing PPP over
satellites brings significant commercial benefits.
Preliminary SBAS test-bed messages and positioning
results have been reported in [1, 7, 8]. Under good satellite
visibility conditions, accuracy at sub-m was obtained when
using SBAS L1 and DFMC, and at sub-decimeter using
PPP. Whilst the PPP positioning method has higher
accuracy than L1 and DFMC SBAS, it requires a period of
This study is funded through project PD8703 SBAS for connected
vehicles part of the Australia/NZ SBAS Test-bed project funded by
FrontierSI on behalf of Geoscience Australia.

convergence time to achieve this level of accuracy, which
can be in the range of 30-60 minutes [9], whereas SBAS L1
and DFMC positioning is achieved almost instantaneously
due to being code-based.
To complement the above studies, this paper presents
performance analysis of the AUS/NZ second generation
SBAS (L1, DFMC) and PPP positioning with a main focus
set on its use for advanced transport applications. The
outcomes of this research can assist in clarifying the
potential of different services of SBAS in improving the
safety and efficiency of the road transport in Australia.
Since transport applications experience continuous changes
in the operation environmental surroundings, the analysis is
conducted classifying the work environment into three
classes: suburban, low -density urban, and high-density
urban. The next section briefly overviews the AUS/NZ
SBAS and discuss processing of data. Next, tests are
described and results are presented and analysed.
Performance metrics such as the obtained precision,
accuracy, availability and suitability of positioning solutions
for ITS applications are next discussed.
II. DESCRIPTION OF THE AUS/NZ SBAS
The second-generation AUS/NZ SBAS test-bed was
developed by Geoscience Australia (GA) and Land
Information New Zealand (LINZ), and is administrated by
the Cooperative Research Centre for Spatial Information
(CRCSI) (now known as FrontierSI). The project was
contracted to GMV (Spain) in conjunction with Lockheed
Martin Space Systems and Inmarsat. The SBAS test-bed
aimed to assess the SBAS performance and benefits over a
diversity of industries including transport, aviation, mining,
agriculture, and maritime applications. The SBAS
infrastructure consists of the space segment, the ground
segment, and the support segment in addition to the user
segment (see Fig 1). The ground segment consists of the
CORS network collecting raw GNSS observations, the
Augmentation Processing Center (APC) computing the
SBAS and PPP corrections, and the uplink system at Uralla
(East Australia) that transforms the SBAS messages into
electromagnetic signals and sends them to the GEO satellite.
The space segment consists of GPS and Galileo satellites,
and the GEO satellite Inmarsat-4F1 (PRN 122) that
transmits the SBAS signals. The support segment is a
magicGNSS Web Monitor at the GMV’s premises in Spain
which monitors the system and service performance [7, 10].
As mentioned earlier, the test-bed continuously
broadcasts the following services [1]:
i- SBAS L1 Legacy service: SBAS broadcasts this signal in
accordance with RTCA/DO229E [11]. The message
broadcasts both corrections and integrity information for
GPS satellite ephemeris, and regional ionosphere
corrections. Since the service is configured for test purposes
only (i.e. not for use in safety applications during the test-
bed period), it broadcast MT 0 message every 6 seconds.
The Test-bed service region is defined to cover the Australia
and New Zealand territories as shown in Figure 2 (left
panel) where the CORS stations used for generation of the
regional ionosphere corrections are located. The ionospheric
delay computed at the reference stations is typically given in
a grid of  (approximately 500km). The user interpolates
these grid values at their location, and computes the slant
ionosphere by using a mapping function.
ii. SBAS L5 DFMC service: The Australian SBAS DFMC
L5 service is implemented in accordance with WG62 GAL
GPS SBAS MOPS v0.3.8_10 draft [3]. The Test-bed signal
contains corrections and DFREI bounds applicable to GPS
L1/L2 + GAL E1/E5a ionosphere-free combinations. The
service coverage area corresponds to the footprint of the
PRN 122 GEO as illustrated in Figure 2 (right panel). In
addition to the larger coverage area, this service is expected
to provide better performance than L1 SBAS during periods
of rapid/high fluctuations of the ionosphere.
iii. Real-time PPP (RT-PPP) through SBAS L1 and SBAS
L5: The AUS/NZ solution uses the spare bits present in the
SBAS message to provide additional information for PPP
with a higher resolution. This is compatible with the SBAS
service, as the SBAS receivers complying with the standard
are instructed to ignore the unused message bits [1].
However, PPP-enabled receivers can use this information to
access the high accuracy data. The PPP corrections
transmitted within the SBAS L1 provides corrections for
GPS L1/L2 signals while the PPP corrections transmitted
through the SBAS L5 signal provide corrections for GPS
L1/L2 + GAL E1/E5a signals. Both services are available in
the entire GEO footprint as shown in Figure 2 (right panel).
In the future operation AUS/NZ SBAS, with enough
number of GPS III satellites become available to allow for
reliable positioning using L1/L5, DFMC SBAS and RT-PPP
will be implemented using the L1/L5 combination instead of
L1/L2 that is used in the test-bed.
Fig. 1. SBAS test-bed infrastructure

Fig. 2. SBAS L1 (left) and SBAS L5 & RT-PPP (right) coverage areas
Whilst many off-the-shelf receivers can pick up L1
SBAS, special hardware and software were needed to
decode DFMC and PPP messages. In our tests, a Septentrio
AsteRx-U receiver and Leica AR10 antenna, with an RF
front-end and Linux tablets were used to capture real-time
signals and log data for post processing. The software
packages developed by GMV, magicGEMINI, which runs
on a Windows platform, was employed for post-processing
SBAS data and magicAPK for decoding messages and
processing RT-PPP. magicGEMINI accepts raw GNSS
observations as well as SBAS messages and outputs SBAS
derived positions and corresponding evaluation of the
system performance at the user level. On the other hand,
magicAPK runs on a Linux platform. Figure 3 depicts a
diagram of the hardware and software setup used.
Fig. 3. Hardware and software setup used
III. SBAS MODELS
The general GNSS code and phase observations can be
expressed as:

󰇛


󰇜

(1)

󰇛


󰇜

(2)
where
P
j
and
φ
j
are the code and phase observables on the
jth frequency (in meters),
and
are the positions of
satellite and user receiver respectively, c is the speed of light
in vacuum,
and
are receiver and satellite clock
offsets and
is the slant ionospheric delay. The term
denotes the zenith tropospheric delay (ZTD) with its
mapping function
. and denote the multipath error on
code and phase measurements respectively,
,
and
,
are the systematic hardware delays, at receiver and
satellite, on code and phase measurements.
is the jth
carrier wavelength, a is the carrier phase integer ambiguity
in cycles, and finally
e
and are the random code and phase
measurements noise, respectively. The unknowns to be
solved for are
and 
. The other terms are sources of
error that need to be accounted for to achieve precise
positioning.
For the SBAS L1 augmented mode, the used receivers
have the option of configuring a carrier phase-based, single-
frequency smoothing filter with individually specified time
constants. No smoothing was applied to the measurements
at the time of recording. Additionally, a proprietary
multipath mitigation technique (APME+, A-Posteriori
Multipath Estimator) is applied to improve measurement
quality by removing short-delay multipath without
introducing biases [1].
In the DFMC solutions, the ionosphere-free combination
is formed using the phase-smoothed code observations on
GPS L1/L2 and Galileo E1/E5a, denoted as

:




󰇛


󰇜

 (3)
where
and
denote the frequencies, and
and
denote
the smoothed code observations on the i and j frequency,
respectively.

denote the noise, where IF is the
ionosphere-free operator. With the satellite orbits and clocks
transmitted as SBAS corrections, and the a priori values of
and
estimated or modelled, the receiver coordinates,
receiver clocks and remaining ZTDs are estimated. The
receiver hardware biases and the common part of the
satellite hardware biases are lumped with the receiver
clocks.
The ionosphere-free code range measurements in the
DFMC SBAS are smoothed by carrier phase measurements
with a Hatch filter that uses a fixed time constant of 100
seconds and a weighting factor (α) as described in
RTCA/DO-253 [12]. The smoothed code observations for
the two frequencies i and j are expressed as:

󰆒






󰇡


󰇢

(4)


 󰇡


󰇢 󰇛 󰇜

󰆒
 (5)
where 

is the ionosphere-free dual-frequency carrier-
smoothed pseudorange at the epoch k,
and
are the
raw pseudorange for frequencies i and j,
and
are the
carrier phase measurement for the two frequencies in cycles
at k, is a frequency ratio, which is (154/115)
2
for Galileo
measurements on E1 and E5a, and (154/120)
2
for GPS
measurements on L1 to L2. In the first 100 seconds since
filter initialization, α is equal to the sample interval divided
by the time since filter initialization; after 100 seconds α is
equal to the sample interval divided by 100 seconds [1].
Newer versions of the DFMC MOPS suggest an increase in
the 100 seconds time constant since the code-carrier
divergence effect is not a worry for the smoothed
ionosphere-free code observations.

For the PPP solutions, the ionosphere-free phase
observations were used in addition to the code observations.
The ambiguity parameters and the ZTDs were modelled as
random walk bias parameters. [4, 13 and 6] describe the
models used in real-time PPP processing.
IV. TEST DESCRIPTION
Testing of the second generation SBAS in this study was
conducted with focus on two focus areas: i ‘Heavy Vehicle
Efficiency’ to examine the SBAS-based positioning ability
to improve transport network efficiency by identifying
which lane the vehicle is travelling in, and ii. ‘Road Safety
to examine the ability of SBAS in supporting issuing
collision alerts to drivers. The tests were carried out in July
and August 2018 in Wollongong and Sydney, Australia. All
data were collected and processed at a frequency of 1 Hz.
One of the vehicles used in these tests is displayed in Figure
3, showing the SBAS antennae mounted on the roof.
Fig. 3. One Test vehicle
As mentioned earlier, it was observed that both SBAS and
PPP positioning performance is heavily dependent on the
environment of application, which varies considerably in the
transport sector. Therefore, the test environment was
subdivided into the following three categories:
Suburban: characterised by low-rise buildings, with a
maximum of three floors,
Low-density urban: with the presence of low-rise and
some high-rise buildings on one side of the road,
High-density urban: characterised by high-rise
buildings on most sides of the road.
An example of one test run that includes all these
environments is illustrated in Figure 4.
To evaluate the accuracy of L1 and DFMC SBAS and
PPP positioning results, a ‘ground truth’ for the positions of
the vehicle was computed in post-processing mode through
independent relative kinematic positioning (PPK). The same
raw code and phase observations used for SBAS-based
positioning were utilized with data from Continuously
Operating Reference Stations (CORS), serving as base
stations, where the test vehicles were within a radius of
eight kilometers from the base stations. Only ambiguity-
fixed solutions from PPK, with 1-5cm precision, were used
as ground truth.
Fig. 4: Test route with subdivision in environments
V. POSITIONING RESULTS
Restricting our focus on transport, where horizontal
positioning is of interest, test results are illustrated in terms
of the time series of the horizontal positioning errors in
North and East directions for the different SBAS positioning
modes for selected representative examples of the tests that
were carried out in this experiment. In addition, the
Horizontal Dilution of Precision (HDOP) is depicted, which
signifies the number and geometry of observed satellites.
The performance indicators considered are availability,
mean error and root mean square error (RMSE), where:
Availability is the fraction of time in which a
position solution was delivered over the total time.
Mean Error and RMSE were computed after all
outliers were removed. The mean error is a measure
of any bias possibly affecting positioning, whereas
the RMSE is a measure of the positioning accuracy.
A. Suburban environment
Figure 5 to Figure 7 depict the L1, DFMC SBAS and
real-time PPP North and East positioning errors respectively
for one example test in the suburban environment. The gaps
seen for some positioning modes refer to unavailable
positions due to observing a limited number of satellites
resulting from signal obstruction by structures or trees.
High-density
Urban
Low-density Urban
Suburban

Fig. 5. SBAS L1 results, suburban environment.
Fig. 6. DFMC results, suburban environment.
Fig. 7. PPP results, suburban environment
From the Figures 5 to 7, one can see that the positioning
performance as expected is highly correlated with HDOP.
For real-time PPP, the figures show that the loss of visibility
of some satellites led to a temporary/partial loss of
convergence, and some time was needed to re-converge to
the previous precision levels. The average statistics for the
full day of testing are given in Table I, listing also for
comparison purpose the results of the traditional Single
Point Positioning (SPP). The table shows positioning
availability, mean error and RMSE. In addition, Figure 8
shows the mean absolute 2D horizontal error for the
different positioning modes, which demonstrates average
positioning errors and hence the system accuracy, regardless
of the sign of the errors. It can be seen that both L1 and
DFMC SBAS gave much better positioning accuracy
compared with SPP, and thus they can provide sub-meter
horizontal positioning accuracy for road applications when
sufficient number of satellites can be observed. The table
and figure 8 show that the overall performance in terms of
availability and accuracy of DFMC SBAS and L1 SBAS is
close, with some improvement in the former compared with
the latter. This is due to the fact that the latter approach uses
interpolated values for the ionosphere delay (introducing
some interpolation errors), while the former removed the
most significant first-order ionosphere delay, at the expense
of increased observation noise, through the use of the
ionosphere-free combination. This trade-off resulted
sometimes in one method giving slightly better results than
the other.
Naturally, PPP after solution convergence, and since it
relays on the more precise carrier-phase observations,
outperformed the L1 and DFMC SBAS methods, which use
carrier phase smoothed-code observations, in terms of all
studied metrics, with 0.2 m PPP average accuracy. The
suitability of the obtained performance and accuracy for
lane level identification, which is a main objective of using
SBAS positioning for ITS applications is discussed in the
next section.
TABLE I - STATISTICS FOR TESTING IN SUBURBAN ENVIRONMENT
Fig. 8. Overall mean horizontal error for suburban test
B. Low-density and High-density urban environments
Figure 9 and Figure 11 show the time series of the
positioning error versus HDOP for DFMC SBAS and PPP
when testing the new generation SBAS in the low-density
urban environment. For the high-density urban areas, the
availability of positioning was below 30% for all
KPI
SPP
SBAS L1
DFMC
Availability
0.96
0.99
0.99
Mean error North (m)
-0.90
0.00
-0.19
Mean error East (m)
-0.08
0.15
0.04
Mean error Up (m)
-1.66
-1.32
-0.45
RMSE North (m)
1.48
0.68
0.57
RMSE East (m)
0.78
0.50
0.32
RMSE Up (m)
2.69
1.75
1.24

Citations
More filters
References
More filters
01 Jan 2003
TL;DR: The International GNSS Service (IGS) provides precise GPS orbit products to the geodetic community with increased precision and timeliness as mentioned in this paper, which can be used to estimate the user position relative to one or multiple reference stations, using differenced carr ier phase observations and a baseline or network estimation approach.
Abstract: Since 1994, the International GNSS Service (IGS) ha s provided precise GPS orbit products to the scient ific community with increased precision and timeliness. Many national geodetic agencies and GPS users interested in geodetic positioning have adopted the IGS precise orbits to achieve centimeter level acc uracy and ensure long-term reference frame stability. Rel ative positioning approaches that require the combination of observations from a minimum of two GPS receivers, with at least one occupying a station with known coordinates are commonly used. The user position can then be estimated relative to one or multiple reference stations, using differenced carr ier phase observations and a baseline or network estimation approach. Differencing observations is a popular way to eliminate common GPS satellite an d receiver clock errors. Baseline or network process ing is effective in connecting the user position to the coordinates of the reference stations while the pre cise orbit virtually eliminates the errors introduc ed by the GPS space segment. One drawback is the practical constraint imposed by the requirement that simultaneo us observations be made at reference stations. An alte rnative post-processing approach uses un-difference d dual-frequency pseudorange and carrier phase observations along with IGS precise orbit products, for stand-alone precise geodetic point positioning (sta tic or kinematic) with centimeter precision. This is possible if one takes advantage of the satellite cl ock estimates available with the satellite coordina tes in the IGS precise orbit/clock products and models systema tic effects that cause centimeter variations in the satellite to user range. Furthermore, station tropo spheric zenith path delays with mm precision and GPS receiver clock estimates precise to 0.03 nanosecond are also obtained. To achieve the highest accuracy and consistency, users must also implement the GNSS-specific conventions and models adopted by the IGS. This paper describes both post-processing approaches, summarizes the adjustment procedure and specifies the Earth and space based models and conventions th at must be implemented to achieve mm-cm level positioning, tropospheric zenith path delay and clo ck solutions. The International GNSS Service (IGS), formerly the International GPS Service, is a voluntary collabora tion of more than 200 contributing organizations in more than 80 countries. The IGS global tracking network of more than 300 permanent, continuously-operating GPS stations provides a rich data set to the IGS Analy sis Centers, which formulate precise products such as s atellite ephemerides and clock solutions. IGS Data Centers freely provide all IGS data and products fo r the benefit of any investigator. This paper focus es on the advantages and usage of the IGS precise orbits and clocks. Currently, up to eight IGS Analysis Centers (AC) contribute daily Ultra-rapid, Rapid and Final GPS orb it and clock solutions to the IGS combinations. The da ily computation of global precise GPS orbits and clocks by IGS, with centimeter precision, facilitat es a direct link within a globally integrated, refe rence frame which is consistent with the current Internat ional Terrestrial Reference Frame (ITRF). Since 200 0 the ultra-rapid product originally designed to serve me teorological applications and support Low Earth Orbiter (LEO) missions, has been made available. The ultra-rapid product has since become useful to many other real-time and near real-time users, as well. For mo re information on the IGS combined solution product s and their availability see the IGS Central Bureau ( see http://www.igs.org/components/prods.html).

576 citations

Journal ArticleDOI
TL;DR: A method is presented to maintain real-time PPP with 3D accuracy less than a decimeter when a sudden communication break takes place and Evaluation of the proposed method in static and kinematic testing shows that positioning precision can be maintained for up to 2 h after the break.
Abstract: The precise point positioning (PPP) is a popular positioning technique that is dependent on the use of precise orbits and clock corrections. One serious problem for real-time PPP applications such as natural hazard early warning systems and hydrographic surveying is when a sudden communication break takes place resulting in a discontinuity in receiving these orbit and clock corrections for a period that may extend from a few minutes to hours. A method is presented to maintain real-time PPP with 3D accuracy less than a decimeter when such a break takes place. We focus on the open-access International GNSS Service (IGS) real-time service (RTS) products and propose predicting the precise orbit and clock corrections as time series. For a short corrections outage of a few minutes, we predict the IGS-RTS orbits using a high-order polynomial, and for longer outages up to 3 h, the most recent IGS ultra-rapid orbits are used. The IGS-RTS clock corrections are predicted using a second-order polynomial and sinusoidal terms. The model parameters are estimated sequentially using a sliding time window such that they are available when needed. The prediction model of the clock correction is built based on the analysis of their properties, including their temporal behavior and stability. Evaluation of the proposed method in static and kinematic testing shows that positioning precision of less than 10 cm can be maintained for up to 2 h after the break. When PPP re-initialization is needed during the break, the solution convergence time increases; however, positioning precision remains less than a decimeter after convergence.

80 citations

24 Sep 2004
TL;DR: In this article, the authors presented the availability of some post-mission precise orbit/clock products for public access and the performance of PPP in static and kinematic modes using single and dual frequency observations.
Abstract: The timely availability of global precise GPS satellite orbit and clock products is enabling the development of precise point positioning (PPP) user applications. Based on the processing of un-differenced pseudorange and carrier phase observations from a single GPS receiver, positioning with centimeter to decimeter accuracy can be attained globally. The PPP approach brings great flexibility to GPS field operations, reduces labor and equipment costs, and simplifies operational logistics by eliminating the need for observation differencing and simultaneous tracking at another location. Seamless integration of the survey results into a global reference frame is another advantage over the differential approach. This paper summarizes the availability of some post-mission precise orbit/clock products for public access and the performance of PPP in static and kinematic modes using single and dual frequency observations . Numerical results are presented showing user positioning accuracy in post-mission and in real-time simulations using NRCan’s On-Line PPP Service and P3, a software package developed at the University of Calgary. From the perspective of GPS correction provision, the impact of tracking network coverage and observation types processed on the correction precision is analyzed with a state-space model implementation using wide-area and global tracking networks. Finally, current limitations of GPS correction sources and the PPP approach are summarized and potential areas for further research and development are proposed.

62 citations


"Performance Analysis of Using the N..." refers background in this paper

  • ...The AUS/NZ SBAS additionally broadcast precise satellite orbits and clock corrections to support real-time float-ambiguity Precise Point Positioning (PPP) service that can deliver 5-20cm accuracy [4, 5, 6, 7]....

    [...]

Journal ArticleDOI
TL;DR: An integrated positioning system is presented, where the global navigation satellite systems real-time kinematic (RTK) method was mainly used and results showed that by integrating GNSS RTK, Doppler with IMU + odometer, positioning was available all the time.
Abstract: Reliable continuing positioning is a critical requirement for intelligent transportation systems (ITS). An integrated positioning system is presented, where the global navigation satellite systems (GNSS) real-time kinematic (RTK) method was mainly used. When RTK is not available, positioning was maintained by using Doppler measurements or by low-cost inertial measurement unit (IMU) coupled with vehicle odometer measurements. A new integrity monitoring (IM) method is presented that addresses each positioning mode of the proposed integrated system. Models for the protection levels (PLs) are presented to bound the position error (PE) along the direction of motion of the vehicle and for the cross-track direction. Both direction components are needed, for instance for collision avoidance and for lane identification. The method was assessed through a kinematic test performed in a dense urban environment. Results showed that by integrating GNSS RTK, Doppler with IMU + odometer, positioning was available all the time. For RTK, positioning accuracy was less than a decimetre and the IM availability was 99%, where the PLs bounded the PEs and were less than an alert limit of 1 m. Positioning using Doppler and IMU + odometer measurements bridged RTK breaks but at the sub-meter level accuracy when used for short periods.

43 citations

Proceedings ArticleDOI
28 Sep 2018
TL;DR: The primary objective of this paper is to provide an update on the Testbed service definition, infrastructure status and performances achieved during 2017-2018 system operation.
Abstract: During 2017 and 2018 a second generation satellite positioning augmentation system is being demonstrated in Australia and New Zealand. This prototype system provides Satellite Based Augmentation (SBAS) and Real Time Precise Point Positioning (PPP) capabilities through the SBAS L1 and L5 signals broadcast from the Inmarsat 4F1 geostationary satellite. The Australia and New Zealand SBAS and PPP Testbed is promoted by Geoscience Australia (GA), Land Information New Zealand (LINZ), and the Australia and New Zealand Cooperative Research Centre for Spatial Information (CRCSI). The system has been developed in collaboration with industry partners Lockheed Martin, Inmarsat and GMV. The services and signals broadcast during the Testbed operations consist of the following: • SBAS L1 legacy service available for GPS L1 single-frequency users in Australia and New Zealand. • SBAS DMFC L5 service available for GPS L1/L2 + GAL E1/E5a dual-frequency users in the Inmarsat 4F1 coverage footprint. • PPP corrections through SBAS L1 message, targeting GPS L1/L2 dual-frequency users • PPP corrections through SBAS L5 message, targeting GPS L1/L2 + GAL E1/E5a dual-frequency users. The transmissions of SBAS signals started in May 2017, while the transmission of the PPP services were initiated in October 2017. The primary objective of this paper is to provide, from the point of view of the system developers, an update on the Testbed service definition, infrastructure status and performances achieved during 2017-2018 system operation.

16 citations

Frequently Asked Questions (1)
Q1. What have the authors contributed in "Performance analysis of using the next generation australian sbas with precise point positioning capability for intelligent transport systems" ?

In this paper, the next generation SBAS and its models are first presented, and the benefits of using its new components are discussed. As expected, the SBAS-based PPP solutions have shown to give the best positioning precision and accuracy among all tested solution types, with sub-decimeter level accuracy, provided that enough convergence time is available. The paper concluded by giving remarks on the use of this new technology for ITS.