JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102, NO. B3, PAGES 5005-5017, MARCH 10, 1997
Precise point positioning for the efficient and robust
analysis of GPS data from large networks
J. F. Zumberge, M. B. Heftin, D.C. Jefferson, M. M. Watkins,
and F. H. Webb
Jet Propulsion Laboratory, California Institute of Technology, Pasadena
Abstract. Networks of dozens to hundreds of permanently operating precision
Global Positioning System (GPS) receivers are emerging at spatial scales that range
from 100 to 10 • km. To keep the computational burden associated with the analysis
of such data economically feasible, one approach is to first determine precise GPS
satellite positions and clock corrections from a globally distributed network of GPS
receivers. Then, data from the local network are analyzed by estimating receiver-
specific parameters with receiver-specific data; satellite parameters are held fixed
at their values determined in the global solution. This "precise point positioning"
allows analysis of data from hundreds to thousands of sites every day with 40-
Mfiop computers, with results comparable in quality to the simultaneous analysis
of all data. The reference frames for the global and network solutions can be free of
distortion imposed by erroneous fiducial constraints on any sites.
Introduction
The Global Positioning System (GPS) has emerged
in the 1990s as the space geodetic technique with the
simultaneous virtues of accuracy and economy [e.g.,
Yunck, 1995]. It has been applied to a variety of
geophysical phenomena, including the motion of tec-
tonic plates [Larson and Freymueller, 1995; Argus and
Heftin, 1995]• plate-boundary deformation [Feigl et al.,
1993], motion associated with earthquakes [Blewitt et
al., 1993; Bock et al., 1993], and changing Earth ori-
entation [Herring et al., 1991; Lindqwister et al., 1992]
and rotation rate. More recent uses of GPS include
volcano monitoring [Webb et al., 1995], ground-based
measurements for atmospheric [Businger et al., 1996]
and ionospheric [e.g., Wilson et al., 1995] applications,
as well as atmosphere and ionosphere sounding [Mel-
bourne, 1995] using low-Earth orbiters equipped with
GPS receivers.
In heavily populated regions of significant seismic ac-
tivity, deployment of dense networks of precision GPS
receivers is in progress. Already hundreds of receivers
are in operation in Japan and dozens in southern Cali-
fornia. Networks such as these allow the on-going mea-
surement of the surface deformation field and are ex-
pected to be valuable both in understanding the com-
plex system of faults in the region and also in hazard
mitigation. Additionally, space-based arrays of GPS-
equipped receivers are expected over the next few years
to exploit the potential applications of GPS to climate,
weather, and the ionosphere.
Copyright 1997 by the American Geophysical Union.
Paper number 96JB03860.
0148-0227/97/96 JB-03860509.00
The volume of GPS data is growing rapidly, and a
means to analyze this volume in a consistent, robust,
and economical manner is essential. In this article we
first discuss the computational burden associated with
processing GPS data in the 'context of data volume and
number of parameters estimated. We show that the
computational burden associated with the rigorous least
squares analysis of data simultaneously from R receivers
scales with R 3. To the extent that global parameters,
that is, orbits of GPS satellites (expressed in an Earth-
fixed reference frame) and satellite clocks, can be esti-
mated with a subset of the R receivers, then data from
the others can be analyzed one at a time. Various anal-
yses of data from R = 49 receivers are used to quantify
the approximations involved in this technique, which
we call "precise point positioning." The validity of the
technique is also demonstrated based on data from over
102 receivers and 104 station days.
Computational Burden
The two GPS data types, carrier phase (L) and pseu-
dorange (P), measure the receiver-to-transmitter dis-
tance with and without, respectively, an unknown bias.
L is a much less noisy measurement than P, which off-
sets the fact that it requires estimation of a bias term.
Both data types exist at each of two frequencies, ap-
proximately 1.2 and 1.6 GHz; this allows the forma-
tion of the ionosphere-free linear combination for each,
which to first order is independent of the ionospheric
electron density. We assume in what follows that we
have formed this combination for L and P.
A phase measurement at time t between receiver r
and transmitter x is modeled as
Lrxt - Prxt + b•t + z•tm(O•t) (1)
+ •,•t + C,•t - c•t +
5005
剩余12页未读，继续阅读
资源评论
BDS_y2016-11-05不要下载 没有什么东西
