24
Geotechnical News • September 2017
GEOTECHNICAL INSTRUMENTATION NEWS
GLONASS (Russia), BeiDou (China)
and Galileo (EU) have been devel-
oped. A GPS satellite is sending three
legacy binary codes known as the
Precise code (P (Y) code), the Coarse/
Acquisition (C/A) code, and the
Navigation (NAV) code. These codes
are modulated into electromagnetic
waves known as L1 at 1575.42 MHz
and L2 at 1227.60 MHz. Both the
code and the carrier signal can be used
for ranging. The code based ranging is
achieved by comparing the time shift
between a section of code from the
satellite and the same synchronized
code generated at the receiver. The
carrier-based ranging requires resolv-
ing the integer number of wavelengths
included in the entire carrier signal
from the satellite to the receiver (inte-
ger ambiguity), which involves more
sophisticated algorithms and yields
more accurate results.
GNSS errors and differential
positioning
Errors exist in all kinds of measure-
ments including GNSS. The main
contributing sources of GNSS errors
are: satellite clock error, satellite
orbit error, ionospheric delay, tropo-
spheric delay, multipath and receiver
noise, causing errors in the orders of
magnitude from a few decimeters to
several meters. Without removing
these errors, the accuracy of GNSS
positioning would not satisfy many
applications including geotechnical
ground movement monitoring, for
which sub-centimeter accuracy is
expected. The solution to eliminate
these errors is differential position-
ing, on which most, if not all, accurate
GNSS positioning techniques rely. In
differential positioning, the position of
a fixed GNSS receiver (referred to as
a base station) is determined to a high
degree of accuracy using conventional
surveying methods. The position of
the base station is also calculated by
using either code-base or carried-base
ranging, which includes the errors
listed above. Because most of the
GNSS errors are spatially related, the
difference between accurate and calcu-
lated position are nearly equal within
a limited geographical area. Therefore,
a spatially close receiver with its posi-
tion in question (rover) can integrate
the ‘difference’ received from base
station via a wireless data link to ‘cor-
rect’ its calculated positon. The closer
the rover is to the base station, the bet-
ter the correction at base could match
to the rover. DGPS (differential GPS)
and RTK (Real-Time Kinematic) are
the two common differential position-
ing techniques. The DGPS is code-
base ranging with 100-200 km typical
baseline (the distance between the
base and the rover), providing approx-
imately +/-1meter accuracy whereas
RTK is carrier phase-based ranging
with 10-20 km baseline, providing cm
level accuracy even when positioning
fast moving objects.
GNSS in geotechnical instru-
mentation and monitoring
Ground/earth structural surface
deformation is one of the most crucial
subjects in geotechnical instrumenta-
tion and monitoring (I&M), for which
GNSS appears to be a perfect tool,
as its direct output is the position of
the object to which the receiver is
attached. Also, there are some unique
advantages of GNSS compared with
other common monitoring methods,
for example: the distance measure-
ment range of GNSS is almost
unlimited in 3D. The base station can
be placed very far away on a stable
zone from the active monitoring
zone. However, GNSS it is still not
commonly considered in geotechni-
cal I&M, mainly due to the follow
reasons:
• It is a less familiar technology to
most geotechnical engineers;
• Its high hardware cost per moni-
tored point (e.g. using high-end
geodetic GNSS receiver );
• Many of the GNSS products are not
capable of delivering millimeter
scale precision;
• Relatively high power needs of the
system to provide near-real time
data (meaning bulky power supply
equipment).
Although there are certain demand-
ing requirements by geotechnical
I&M, we shouldn’t neglect there are
also some very ‘favorable’ conditions
compared with other GNSS applica-
tions when designing a GNSS-based
monitoring system:
• Although a moving rate of centi-
meters per day is quite significant
to geotechnical engineers, it is
still considered ‘static’ positioning
for GNSS which was originally
designed to track fast moving
objects;
• The area of the monitoring zone is
usually not large, so the base sta-
tion can be located closely to the
rover (< 5km), which will help to
improve the accuracy of differen-
tial positioning;
• The monitoring data is usually
only required to be updated every
few hours or even less frequently,
while the sampling rate of GNSS
is usually in ‘Hz’.
Implementation of GNSS to
monitor landslide movement
A recent pilot project performed by
Sixense and Washington State Depart-
ment of Transportation (WSDOT)
geotechnical office has implemented a
GNSS system in a small landslide site
in Washington State. The project site is
located along a short section of a noto-
riously unstable 40 km long stretch of
US Highway 101, between the cities
of Aberdeen and Raymond, which
is about 170 km to the south east of
Seattle (Site photos are shown in Fig.
1 and Fig. 2). This site suffers from
frequent small-scale landslide move-
ment, especially during the Pacific
Northwest rainy season (November –
April). The active landslide head scarp
is estimated to be about 100 meters
in length and the presumed landslide
toe is around 175 meters downslope
(~2H:1V slope), near an un-named
creek. Over the last decade, WSDOT
maintenance crews have had to resur-
face the highway on an annual basis
