Differential GPS - Understanding GPS: Principles and Applications - page 418

N NDD
l

λ −

wl wl cp

= (8.44)

l 2

1

λ

l 1

Care must be taken at this point, since the calculation of the L1 ambiguity set
will occasionally be incorrect. Referring to (8.14), it is shown that the carrier phase
DD also contains an amount of noise; ultimately, this noise is swept into the resolved
ambiguities. An intuitive glance at (8.44) leads to the conclusion that conversion to
the L1 ambiguity set seldom if ever produces integer values. Generally speaking, the
results are very close to integers, and the proper set can usually be realized by pick-
ing the nearest integer values. Occasionally, however, there is enough noise on one
or more of the wide-lane measurements to cause the next higher or lower integer
ambiguity value to emerge from the conversion process. Reference [20] uses
wide-lane techniques with subsequent conversion to the L1 wavelength for ambigu-
ity resolution and points out that a phase error as small as 1.2 cm can produce a con-
version error of 9.72 cm ( /2 at L1), which results in the selection of the wrong

ambiguity if the nearest integer is chosen. The conclusion is that, while reversion to
single-frequency tracking adds robustness, the conversion process must be done
with care. The L1 integer-ambiguity values that are generated by rounding the
results from (8.44) must be near integer values to begin with or the operation poten-
tially becomes suspect. One approach to solving this problem would be to follow
(8.44) with a limited search around the L1 ambiguities.
As a final note, starting in 2007, GPS satellites will incorporate an L5 signal at
1,176.45 MHz. This signal will permit the construction of an extra-wide-lane metric
from the difference between L2 and L5. The resulting wavelength will be 5.861m,
and extremely rapid ambiguity searches will result. With the reliability of the cur-
rent GPS constellation, full operational capability for use of the L5 signal will proba-
bly not occur before 2015.

λ

8.4.2 Static Application
While land surveying is probably the most common of static applications, many
other near-static applications are taking advantage of interferometric techniques.
Among these could be counted precise dredging requirements for harbors and
inland waterways, accurate leveling of land for highway construction and agricul-
tural needs (especially land under irrigation), trackage surveys done to exacting
standards for high-speed rail service, and a whole host of others. Generally, the driv-
ing factor in near-static or low-dynamic applications is the necessity for centime-
ter-level accuracy in the vertical dimension. For land surveying, requirements for
accuracies in the millimeter regime in three dimensions are not uncommon. The
classical approach, used initially in [16], demanded occupation times of up to sev-
eral hours with simultaneous collection of GPS pseudorange and phase data at both
ends of a prescribed baseline. This classic paper reported that “analyses of data from
different observation periods yielded baseline determinations consistent within less
than 1 cm in all vector components.” That was in December 1980, the baseline was
92.07m, and the occupation time was a minimum of 1 hour. The survey data was
processed after the fact, as remains typical today. The requirement for the occupa-