没有合适的资源？快使用搜索试试~ 我知道了~
温馨提示
试读
34页
【2015Kouba】A GUIDE TO USING INTERNATIONAL GNSS SERVICE (IGS) PRODUCTS 原文，可能有的朋友找不到，特此分享一下。 文章摘要如下： Since 1994, the International GNSS Service (IGS) has provided precise GPS orbit products to the scientific community with increased precision and timeliness. Many national geodetic agencies and GNSS (Global Navigation Satellite System) users interested in geodetic positioning have adopted the IGS precise orbits to achieve centimeter level accuracy and ensure longterm reference frame stability.
资源推荐
资源详情
资源评论
A GUIDE TO USING INTERNATIONAL GNSS SERVICE (IGS)
PRODUCTS
Jan Kouba
Geodetic Survey Division
Natural Resources Canada
615 Booth Street, Ottawa, Ontario K1A 0E9
Email: koub[email protected]a
Updated September 2015
Abstract
Since 1994, the International GNSS Service (IGS) has provided precise GPS orbit products to the scientific
community with increased precision and timeliness. Many national geodetic agencies and GNSS (Global
Navigation Satellite System) users interested in geodetic positioning have adopted the IGS precise orbits to
achieve centimeter level accuracy and ensure longterm reference frame stability. Relative positioning
approaches that require the combination of observations from a minimum of two GNSS 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 carrier phase observations and a
baseline or network estimation approach. Differencing observations is a popular way to eliminate
common GNSS satellite and receiver clock errors. Baseline or network processing is effective in
connecting the user position to the coordinates of the reference stations while the precise orbit virtually
eliminates the errors introduced by the GNSS space segment. One drawback is the practical constraint
imposed by the requirement that simultaneous observations be made at reference stations. An alternative
postprocessing approach uses undifferenced dualfrequency pseudorange and carrier phase observations
along with IGS precise orbit products, for standalone precise geodetic point positioning (static or
kinematic) with centimeter precision. This is possible if one takes advantage of the satellite clock estimates
available with the satellite coordinates in the IGS precise orbit/clock products and models systematic
effects that cause centimeter variations in the satellite to user range. Furthermore, station tropospheric
zenith path delays with mm precision and GNSS 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 postprocessing
approaches, summarizes the adjustment procedure and specifies the Earth and space based models and
conventions that must be implemented to achieve mmcm level positioning, tropospheric zenith path delay
and clock solutions.
1. Introduction
The International GNSS Service (IGS) (Dow et al., 2009), formerly the International GPS Service, is a
voluntary collaboration of more than 200 contributing organizations in more than 80 countries. The IGS
global tracking network of more than 300 permanent, continuouslyoperating GNSS stations provides a
rich data set to the IGS Analysis Centers, which formulate precise products such as satellite ephemerides
and clock solutions. IGS Data Centers freely provide all IGS data and products for the benefit of any
investigator. This paper focuses on the advantages and usage of the IGS precise orbits and clocks.
Currently, up to ten IGS Analysis Centers (AC) contribute daily Ultrarapid, Rapid and Final
GPS/GLONASS orbit and clock solutions to the IGS combinations. The daily computation of global
precise GPS/GLONASS orbits and clocks by IGS, with centimeter precision, facilitates a direct link within
a globally integrated, reference frame which is consistent with the current International Terrestrial
Reference Frame (ITRF). Since 2000 the ultrarapid product originally designed to serve meteorological
applications and support Low Earth Orbiter (LEO) missions, has been made available. After several years
of pilot project operation, on April 1, 2013, the IGS Real Time Service (RTS) has been officially launched,
providing precise orbit/clocks products in real time. The ultrarapid and RTS products have since become
useful to many other realtime and near realtime users, as well. For more information on the IGS combined
solution products and their availability see the IGS Central Bureau (see http://igs.org/products).
For GNSS users interested in meter level positioning and navigation, a simple point positioning interface
combining pseudorange data with IGS precise orbits and clocks (given at 15 min intervals) can be used
(e.g. Héroux et al., 1993; Héroux and Kouba, 1995). Since May 2, 2000 when GPS Selective Availability
(SA) was switched off these products also satisfy GPS users observing at high data rates in either static or
kinematic modes for applications requiring meter precision. This is so, because the interpolation of the 15
min SAfree satellite clocks given in IGS sp3 files is possible at the precision level of a few dm.
Furthermore, since December 26, 1999, separate, yet consistent, clock files, containing separate
combinations of satellite/station clocks at 5min sampling intervals have been available and on November
5, 2000, the clock combinations became the official IGS clock products (Kouba and Springer, 2001). The
5min clock sampling allows an interpolation of SAfree satellite clocks well below the dm level
(Zumberge and Gendt, 2000). In order to keep clock interpolation errors at or below the cmlevel, starting
with GPS Week 1410 (January 14, 2007), the IGS Final clock combinations also include additional clock
files with 30sec sampling. For GNSS users seeking to achieve geodetic precision, sophisticated processing
software packages such as GIPSY (Lichten et al., 1995), BERNESE (Dach et al., 2007) and GAMIT (King
and Bock, 1999) are required. However, by using the IGS precise orbit products and combining the GNSS
carrier phase data with nearby IGS station observations, geodetic users can achieve precise relative
positioning consistent with the current global International Terrestrial Reference Frame (ITRF), with great
ease and efficiency and with relatively simple software. For example, differential software packages
provided by receiver manufacturers may also be used, as long as they allow for the input of the station data
and orbit products in standard (IGS) formats and conform to the international (IGS and IERS) conventions
and standards (see Section 5.3).
The precise point positioning (PPP) algorithms based on undifferenced carrier phase observations have
been added to software suites using undifferenced observations such as GIPSY (Zumberge et al., 1997)
and more recently even the traditional doubledifferencing software package such as the BERNESE has
been enhanced also to allow precise point positioning. Users now have the option of processing data from
a single station to obtain positions with centimeter precision within the reference frame provided by the
IGS orbit products. PPP eliminates the need to acquire simultaneous tracking data from a reference (base)
station or a network of stations. It has given rise to centralized geodetic positioning services that require
from the user a simple submission of a request and a valid GNSS observation file (see e.g., GhoddousiFard
and Dare, 2005). An alternative approach is an implementation of simple PPP software that effectively
distributes processing by providing portable software that can be used on a personal computer. This
software then takes full advantage of consistent conventional modeling and the highly accurate global
reference frame, which is made available through the IGS orbit/clock combined products.
For both relative and PPP methods that utilize IGS orbit/clock products, there is no need for large and
sophisticated global analyses with complex and sophisticated software. This is so because the IGS
orbit/clock products retain all the necessary information of the global analyses that have already been done
by the IGS ACs, using the state of art knowledge and software tools. Thus, the users of the IGS products in
fact take full advantage of the IGS AC global analyses, properly combined and quality checked, all in
accordance with the current international conventions and standards.
2. IGS Orbit/Clock Combined Products
Even though, strictly speaking, it is illegitimate to combine solutions that are based on the same
observation data set, the combinations of Earth Rotation Parameters (ERP) and station coordinate solution
submissions have been successfully used by the International Earth Rotation and Reference Systems
Service (IERS) for many years. Such combinations typically result in more robust and precise solutions,
since space technique solutions are quite complex, involving different approaches and modeling that
typically generate a randomlike noise which is then averaged out within the combination process. This
approach is also valid for the combination of IGS orbit solutions as clearly demonstrated by Beutler et al.,
(1995) who have also shown that, under certain conditions, such orbit combinations represent physically
meaningful orbits as they still satisfy the equations of motions. Furthermore, when the AC weights
resulting from orbit combinations are used in the corresponding ERP combinations (as done for all IGS
combinations before February 27, 2000 and currently still being done for the Rapid and UltraRapid ones
only), the crucial consistency between the separately combined orbits and ERP solutions is maintained.
The IGS combined orbit/clock products come in various flavors, from the Final, Rapid, the UltraRapid to
RTS ones. The IGS UltraRapid (IGU) products replaced the former IGS predicted (IGP) orbit products on
November 5, 2000 (GPS Week 1087, MJD 5183), the RTS combined products have been officially made
available since April 1, 2013 (GPS Week 1786/Day 2, MJD 56748). The IGS combined orbit/clock
products differ mainly by their varying latency and the extent of the tracking network used for their
computations. The IGS Final orbits (clocks) are currently combined from up to nine contributing IGS
ACs, using seven, largely independent, software packages (i.e. BERNESE, GAMIT, GIPSY, NAPEOS
(Springer et al., 2011), EPOS (Gendt et al., 1999), PAGES (Schenewerk et al., 1999) and GINS/Dynamo
(Marty 2009). The IGS Final orbit/clocks are usually available before the thirteenth day after the last
observation. The Rapid orbit/clock product is combined 17 hours after the end of the day of interest. The
latency is mainly due to the varying availability of tracking data from stations of the IGS global tracking
network, which use a variety of data acquisition and communication schemes, as well as different levels of
quality control. In the past, the IGS products have been based only on a daily model that required
submissions of files containing tracking data for 24hour periods. In 2000, Data Centers have been asked
to forward hourly tracking data to accelerate product delivery. This new submission scheme was required
for the creation of an UltraRapid product, with a latency of only a few hours (currently 3 hours), which
should satisfy the more demanding needs of most realtime users, including the meteorological community
and LEO (Low Earth Orbiter) missions. It is expected that IGS products will continue to be delivered with
increased timeliness in the future (Weber et al., 2002a). Development of true realtime products, mostly
satellite clock corrections, has been accomplished by the IGS RealTime Pilot Project, which on April 1,
2013 has resulted in the IGS RTS (Caissy et al., 2012). For more information on the IGS products and their
possible applications see e.g. Neilan et al. (1997); Kouba et al. (1998b) and Dow et al. (2005).
Figure 1: Weighted orbit RMS of the IGS Rapid (IGR) products and AC Final orbit solutions during 1994
2015 with respect to the IGS Final orbit products. (COD – Center for Orbit Determination in Europe,
Switzerland; EMR – Natural Resources Canada; ESA  European Space Agency; GFZ –
GeoForschungsZentrum Potsdam, Germany; GRG  Groupe de Recherche de Géodésie Spatiale (GRGS)/
Centre National d'Etudes Spatiales (CNES)
;
JPL – Jet Propulsion Laboratory, U.S.A.; MIT Massachusetts
Institute of Technology, U.S.A.; NGS – National Geodetic Survey, NOAA, U.S.A.; SIO  Scripps Institute
of Oceanography, U.S.A.). (Courtesy of the IGS ACC, see
http://acc.igs.org/)
From Figure 1, one can see that over the past 15 years the precision of the AC Final orbits has improved
from about 30 cm to about 1  2 cm, with a concomitant improvement in the IGS Final combined orbit. The
current precision of IGS combined orbits has been attained since about GPS Week 1400 ( November 5,
2006) when the IGS05 realization of ITRF2005 has been adopted. It is also interesting to note that the IGS
Rapid orbit combined product (IGR), with less tracking stations and faster delivery times, is now more
precise than the best AC Final solutions. The precision of the corresponding AC/IGS ERP solutions has
shown similar improvements since 1994. One element that has received less attention is the quality of the
GPS satellite clock estimates included in the IGS orbit products since 1995. Examining the summary plots
for IGS Final clock combinations at the IGS AC Coordinator (ACC) web site (http://acc.igs.org/), one can
notice that after small biases are removed, the satellite clock estimates produced by different ACs now
agree with standard deviations of 0.02  0.06 nanosecond (ns) or 1  2 cm. This is also consistent with the
orbit precision. Any biases in the individual IGS satellite clocks will be absorbed into the phase ambiguity
parameters that users must estimate. The precise GNSS orbits and clocks, weighted according to their
corresponding precision (sigmas), are the key prerequisites for PPP, given that the proper measurements are
made at the user site and the observation models are implemented correctly.
3. Observation equations
The ionosphericfree combinations of dualfrequency (f
1
, f
2
) GPS (or GNSS) pseudorange (P) and carrier
phase observations (
Φ
) are related to the user position, clock, troposphere and ambiguity parameters
according to the following simplified observation equations:
ℓ
P
=
ρ
+ c(dTdt) + T
r
+
ε
P
( 1 )
φ
ℓ
=
ρ
+ c(dTdt) + T
r
+ N
λ
+
ε
Φ
( 2 )
Where :
ℓ
P
(P3) is the ionospherefree combination of P
1
and P
2
pseudoranges (f
1
2
P
1
 f
2
2
P
2
)/(f
1
2
f
2
2
),
ℓ
Φ
(L3) is the ionospherefree combination of L1 and L2 carrierphases (f
1
2
λ
1
φ
1
 f
2
2
λ
2
φ
2
)/(f
1
2
f
2
2
),
f
1
, f
2
are the L1, L2 frequencies (for GPS 1575.42 and 1227.6 MHz, respectively)
dT is the station receiver clock offset from the GPS (or GNSS) time,
dt is the satellite clock offset from the GPS (or GNSS) time,
c is the vacuum speed of light,
T
r
is the signal path delay due to the neutralatmosphere (primarily the troposphere),
N is the noninteger ambiguity of the carrierphase ionospherefree combination,
λ
1,
λ
2
, λ are the of the carrier phase L1, L2 and L3combined (10.7cm) wavelengths, respectively,
ε
P
,
ε
Φ
are the relevant measurement noise components, including multipath.
Symbol
ρ
is the geometrical range computed by iteration from the satellite position (Xs, Ys, Zs) at the
transmission epoch t and the station position (x, y, z) at the reception epoch T = t +
ρ
/c, i.e.
)()()(
222
zZsyYsxXs −−−
++=
ρ
.
Alternatively, for relative positioning between two stations (i, j) the satellite clock errors dt can be
eliminated simply by subtracting the corresponding observation Eqs. (1) and (2) made from the two stations
(i, j ) to the same satellite (k), i.e.:
ℓ
Pij
k
=
∆ρ
ij
k
+ c
∆
dT
ij
+
∆
T
rij
k
+
∆ε
Pij
k
,
(3)
ℓ
Φ
ij
k
=
∆ρ
ij
k
+ c
∆
dT
ij
+
∆
T
rij
k
+
∆
N
ij
k
λ
+
∆ε
Φ
ij
k
,
(4)
where
∆
(.)
ij
k
denotes the single difference. By subtracting the observation Eqs. (3) and (4) pertaining to the
stations (i, j) and the satellite k from the corresponding equations of the stations (i, j) to the satellite l, we
can form so called double differenced (DD) observation equations, where the station clock difference errors
∆
dT
ij ,
which are the same for both
single differences, are
also eliminated, which is generally true for GPS
where all the satellites use common carrier frequencies:
ℓ
Pij
kl
=
∆ρ
ij
kl
+
∆
T
rij
kl
+
∆ε
Pij
kl
,
(5)
ℓ
Φ
ij
kl
=
∆ρ
ij
kl
+
∆
T
rij
kl
+
∆
N
ij
kl
λ
+
∆ε
Φ
ij
kl
,
(6)
where
∆
(.)
ij
kl
represents the respective double difference for the (i, j) station and (k, l) satellite pairs
.
Furthermore, the initial L1 and L2 phase ambiguities that are used to evaluate
the
ionosphericfree
ambiguities
∆
N
ij
kl
become integers. This is so since the fractional phase initializations on L1 & L2 for the
station (i, j) and satellite (k, l) pairs, much like station/satellite clock errors, are also eliminated by the
above DD scheme. Consequently, once the L1 and L2 ambiguities are resolved, the ionosphericfree
ambiguities
∆
N
ij
kl
become known and can thus be removed from the equation (6), which then becomes
equivalent to the pseudorange equation (5), i.e. double differenced phase observations with fixed
ambiguities become precise pseudorange observations that are derived from unambiguous precise phase
measurement differences. That is why fixed ambiguity solutions yield relative positioning of the highest
possible precision, typically at or below the mm precision level (e.g. HofmannWellenhof et al., 1997).
The above integer phase ambiguity resolution may be compromised for GLONASS, since, unlike GPS and
other GNSS', GLONASS satellites use slightly different carrier frequencies (frequency channels). Then, a
small offset between pseudorange and phase observation epochs, constant and usually known within a
receiver, causes small, frequency channel dependent phase biases (Sleewaegen et al., 2012). However,
when GLONASS data phase and pseudorange observations refer to the same epoch (clock) as required by
the RINEX format, there are no GLONASS inter channel phase biases! If uncorrected (i.e., improperly
RINEXed) GLONASS data is used, DD phase observations no longer have integer character and the
GLONASS channel phase biases are absorbed into the real phase ambiguity solutions. Note that the
GLONASS frequency dependent pseudorange biases, which can reach up to several m, unlike the phase
ones, are real and cannot be removed during RINEXing. They have to be either externally calibrated, or
mitigated by an appropriately increased standard deviation of GLONASS pseudorange observations.
The equations (1), (2) and (5), (6) appear to be quite different, with a different number of unknowns and
different magnitudes of the individual terms. For example, the double differenced tropospheric delay
∆
T
rij
kl
is much smaller than the undifferenced T
r, ,
the noise
∆ε
(.)
ij
kl
is significantly larger than the original, un
differenced noise
ε
(.), etc. Nevertheless, both undifferenced and DD approaches produce identical results,
provided that the same set of undifferenced observations and proper correlation derived from the
differencing, are used. In other words, the DD difference position solutions with properly propagated
observation weight matrix (see e.g. HofmannWellenhof et al., 1997), are completely equivalent to un
correlated, undifferenced solutions where the (satellite/station) clock unknowns are solved for each
observation epoch.
Since we are using the IGS orbit/clock products, the satellite clocks (dt) in Eqs. (1) and (2) can be fixed
(considered known) and thus can be removed. Furthermore, expressing the tropospheric path delay (T
r
) as a
product of the zenith path delay (zpd) and mapping function (M), which relates slant path delay to zpd,
gives the point positioning mathematical model functions for pseudorange and phase observations:
f
P
=
ρ
+ c dT + M zpd +
ε
P

ℓ
P
= 0,
( 7 )
f
Φ
=
ρ
+ c dT + M zpd + N
λ
+
ε
Φ

ℓ
Φ
= 0;
( 8 )
The tropospheric path delay (M zpd ) is separated in a predominant and wellbehaved hydrostatic part (M
h
zpd
h
) and a much smaller and volatile wet part (M
w
zpd
w
). While zpd
h
can be modeled and considered
known, zpd
w
has to be estimated. For most precise solutions, temporarily varying zpd
h
, M
h
and M
w
have to
be based either on global seasonal models (Boehm et al., 2006; Boehm et al., 2007), NWM  numerical
weather models (Boehm and Schuh, 2004; Kouba, 2007), or alternatively (and potentially more precisely)
剩余33页未读，继续阅读
资源评论
 开开心心过大年620240423#完美解决问题 #运行顺畅 #内容详尽 内容十分有用，帮助很大
流浪猪头拯救地球
 粉丝: 8751
 资源: 30
上传资源 快速赚钱
 我的内容管理 展开
 我的资源 快来上传第一个资源
 我的收益 登录查看自己的收益
 我的积分 登录查看自己的积分
 我的C币 登录后查看C币余额
 我的收藏
 我的下载
 下载帮助
安全验证
文档复制为VIP权益，开通VIP直接复制
信息提交成功