Proceedings ofthe IEE/EOES/C¶ktTC Ninth Working Conference on Current f2leasurement Technofogy

Remote sensing stream flow and soil moisture by

using reflected GPS Signals Li & L2 observation and
Doppler shifts with an Integrated GPS Receiver
L. C. Shen', J. C. Juang2, C. L. Tsai'
lEarth Science and Academy Department, NCKU,
2Electronic Machine and Academy Department, NCKU,
No 1. University Rd.
Tai-Nan, Taiwan

Abstract--In this paper, a new application and development environment module could be enhanced by two map datasets
of a highly integrated GPS receiver with reflected GPS signals of the image map, Taiwan TOPO couture map, and Taiwan's
for stream flow and soil moisture will be described. First, both Digital Terrain Evaluation Data. (4) Signal treatment module
Right Hand Circular Polarization (RHCP) and Left Hand . . .
Circular Polarization (LHCP) antennas are employed so that
direct and reflected signals can be acquired simultaneously. The
dtmines rf signal paths ting int accont
combinatons of signal interactons (reflectons and directons)
direction of arrival of the signals may be along the reflected for classification of Ground object detection. (5) The receiver
signal path or even along the line-of-sight of a particular satellite. position module used an altitude iteration loop into least
An integer ambiguity algorithm has also been implemented. The squares searching for receiver position, which was modified
precise point positions for RHCP and LHCP antennas are and enhanced by giving 4 or more pseudo range and
enhanced and processed by repeating instantaneous ionosphere ephemeredes. (6) The 3 dimensional velocity estimated model
delay correct model with deriving from LI and L2 carrier phase
and troposphere estimated parameter model. During the
ephemeredes.s(6) The 3 2dim nsional eate d mel
shall predict stream flow [2] and stream water level in real
development and test stage, the Digital Terrain Elevation Data time by using GPS L1/L2 Doppler shift and reflected stream
(DTED) and visual elements of satellite's images has been used moving surface position.
and mapped with the integrated software. For remote sensing of This paper is structured as follows. In the section II, system
river, ocean, and landscape, the accuracies of each reflected diagram for integrated GPS receiver, GPS reflected positions,
altitude are among 10 cm and 30 cm. The accuracies of each surface roughness, classification of object detection and
reflected area are converged among 2 cm and 10 cm. stream flow estimated methods are presented. In section 'II,
Unlike most existing GPS reflection experiment, The goal The remote sensing stream flow that based on the receiver
of the study is to exploit the carrier phase, Doppler shift,
reflectivity of L1I/L2 SINo signal-to-noise density ratio
remon sensing stream
position and 3 dimensional stream velocity method are briefly
velocit med are rier
components of the reflected signals and direct signals for stream outlined. The stream soil moisture should be classified by
flow and for stream clam water, disturbed water, dry soil, wet saturated water for soil and L1/L2 reflectivity on GPS
soil object detection with surface. The soil moisture should be reflected footprint. Section IV. Finally, conclusions are drawn
classified by volumetric content of saturated water for soil and inspired by confirmation by ongoing experiments of the theory
L1IL2 reflectivity on GPS reflected footprint. The three expounded in this paper.
dimensional stream flow modeling is predicted by using Doppler
shifts due to surface reflection as a moving surface on river. The
each instantaneous moving surface should be exploited by each II. INSTRUMENT AND METHODS
reflected GPS carrier phase and reflected point. A. Basic Principles
The system diagram for remote sensing of stream water
level and soil moisture are shown as Figure. 1. The instrument
I. INTRODUCTION
A reflected GPS signal contains information about the A
reflecting object since the characteristics of the reflected T G
~~~~~~~~~~~~data
to MIATLA4B ASCII
signal varies a lot depending upon the reflecting object [1] by ,,/ RHCP $,,
Down LoadGPS
_
PATfile
format

utilizing GPS-derived classification features in combination

-1~ ~ ~ ~~ ~~~ ~ ~ ~ ~ ~ ~ ~ ~ ....01 rcie
4 o bseti"o data Specify heights~ofHICP and
by PC or NB dLndulation
L-ICP anithh
with visible wavelength imagery. In developing software correcion k,l- ofDTED
process, Six modules almost were written in MATLAB®:(1). DownLoadGPS Compute. RHCPpsitisd
The satellite motion module uses receiver's raw data including poin t2 positions,RpflectiVityfor
almanac and ephemeris data. (2) The reflected points and receiver sfdiftsofPRNfor _ streflov,emated
trajectory module combine and use the results of receiver X IP, olkorfetd nIsu.
position module and satellite motion module. (3) The\
Figure. 1. System diagram for remote sensing of stream flow and stream soil
moisture

1-4244- 1486-5/08/$25.00O©2008 IEEE 181

 Proceedings ofthe IEE/EOES/C¶ktTC Ninth Working Conference on Current f2leasurement Technofogy
is mounted on a 6.5-meter-length of steel pipe with one RHCP
and three LHCP antennas by using two SOKKIA 2600 GPS
L1I/L2 Band receivers. The height of the bridge is 13.2 m
(from the bridge floor to the bridge pier). The altitude of the
bridge floor is 446.25 m with regard to the ground sea level.
Consequently, information pertaining to the water level of the
stream, stream flow, and ground reflective index can be
obtained. Figure 2 depicts a typical experimental setup in
which a 6.5-m boom is extended from a base so that the two
antennas are placed above a stream. The three LHCP antennas
are placed orthogonally on the stream surface so that the
reflected signals from different directions can be observed. In
contrast, only one RHCP antenna is used to receive the direct
GPS signals. An enhancing reflection ray treatment processes
was taken [3], [6] by using N-P connector to combine
reflected signals with three LHCP antenna (Type: H601; Dbic:
9 dB-Hz) connect to one receiver's RF P-N connector and
RHCP antenna (Type: NovAtel GPS-700; Dbic: 7 dB-Hz)
connect to the other one receiver's RF P-N connector. The
GPS Li/L2 Band receivers are two SOKKIA-2600 type
receivers (same of motherboard as NovAtel DL-4).
In this measurement connection, the measurement for
remote sensing of stream water level, soil classification and
ground object detection can be found in a related conference
paper [7], [8] and in [17].
by Directive signals observations or Reflected signals
observations. Sls: the value of c / Nl Carrier-to-noise density
ratio of LI band by Directive signal observations or Reflected
signals observations; S2s: the value of C / N' Carrier-to-noise
2
density ratio of L2 band by Directive signal observations or
Reflected signals observations. The signal-to-noise density
ratio is described as following in SI/N- =lln(CIN'/10)
0 Os/I0
fis:the value of Doppler shift for LI frequency by Directive
signal observations or Reflected signals observations; tihe
s
value of Doppler shift for L2 frequency by Directive signal
observations or Reflected signals observations.
A MATLAB Software simulation tool has been developed that
takes statistical results from the Ground truth measurements of
one or two receivers and maps them against the sky to
determine geodetic images such as satellite ID, receiver
position, reflected area, altitude estimation, GPS clock time
delay At = (dt - dT)t', ionosphere delay I{t and troposphere
delay D' (Was shown as Figure. 3.). The P1, pseudo range
solution of observation point, the floating points shall be
quickly solution of Least square searching method with
z GPS CPS Satellite

H q ot(LL4
Lo ~~~~~~~RHCP GPS,
Xd
'1
1Zd) (l Y Z
iaSignTas
M |x

SeaorGrouind surface

LHCP ounna1d
oniputed potion,-
Figure 2. Antenna setup during data collection at Lon Ann Bridge ADI i hD
The Integrated Direct and Reflected rays were crossing e T-
along the line of sight and reflection path between each
satellite orbit and RHCP antenna point, LHCP antenna point.
The MATLAB tool is built to use GPS observations 11s, )2s, Earth center
Pls, P2s, Sls, S2s , f2tiwihtmtalprcdr.Rcve
tU
with time table procedure. Receiver Figure. 3. The Profile of Reflected Ground surface position and GPS satellite
soon enhances 90% of the receiver capability of reflected position: show as LHCP position CEP: 10-50 cm and RHCP position CEP: 2-10
signals on low altitude of receiver position [9]. . A Least cm were computed by five GPS carrier phase observations from different GPS
. ~~L1/L2 Band Receivers.
squares searching algorithm for receiver position was
enhanced and developed by centimeter level altitude loop. enhancing 10-cm level altitude loop. The Ols, (12s carrier
(Goad, 1997)[10]; s = d: Direct signal; s = r: Reflected signal. phase solution of observation point, a modified receiver
Dls: phase of Li band by Direct signals observations or position searching method was developed and described by
combined altitude iteration integer ambiguity
loop forandMATLAB
Reflected signals observations; 4I42s: phase of L2 band by estimation of ambiguity solution [II]. This
Direct signals observations or Reflected observations. P15: carrier phase solution of residuals was minimization to small
pseudo range of Li band by Directive signals observations or residuals by iteration of observations with enhancing i-cm
Reflected signals observations; P2S: pseudo range of L2 band

182
 Proceedings ofthe IEE/EOES/C¶ktTC Ninth Working Conference on Current f2leasurement Technofogy
level altitude loop Least square searching receiver position There is the way of determining the height of the scattering
process and reflected point's position estimation. [ 12] h of RHCP receiver position for use in height estimation.
B. Ground Height Estimation and Surface Roughness DTED
When each Range of Reflection points and earth center The Digital Terrain Elevation Data (DTED level 2; 30 M
are same value in the same stream surface area, the new per pixel) is correct in an absolute sense, then a simple
refection point (X Y Z) will be located and subtraction of the DTED data of area is sufficient [2][4]. The
accumulated projection line with time. The RHCP and LHCP processing diagram of reflectivity and Ground object detection
fixed position cause different altitude for observation point of simulation program show as Figure. 4. The altitude trial
when the mean of ground altitude is subject to estimated
ground or stream altitude of reflected area per each satellite,
since mean value of RHCP altitude and LHCP altitude start
and product the mean of ground altitude. A reflected point is Cmt
Compu3te:
described for stream surface with slope of line of sight -RHCP position
between satellite and RHCP/LHCP receiver position. The -LHCP position
Direct receiver position (Lo n Lat H ) RHCP shall transfer
into (X d yd z d ) coordinate and the 3-D geometry of
Compute:
-GPS Reflected point position
Iby1 DTED
specif lixr
model (-10 meter
-Generate Foot-print <_accuracy of pixel for
line of sight between Receiver position on Taiwan and five -Gnerate iigitalTerrai
(reflected GPS signals generators satellites (show as Figure. 3). ElevationData)
The Reflected receiver position (Lon Lat H ti). _
tt
transfer into (X Yr z ); ti: instantaneous number of Specify: (Object
r r r
~~~~~~~~~~~~~~~~~~~ ~ ~~~~~Generate
~ and classify- Indzex by RGBE
satellites for LHCP receiver. Satellites position fg,thereee pointh intensityofGATimage
(Lon Lat H)is transfer to (Xis yis z si) i: ID of Object Detection Index (-1 meter accura
pi_xefoSA image)o
satellite; The Ground or water surface altitude is described as
following Eq. (1) As the project of reflected point on line of observationtimeK
IGenlerate the sample vectors of
r:
sight is (Lon Lat H )roud ground transfer to cton ocient Dif(AL (
-Soilreff
(X Y z )i; The Eq. (1) conduct to Radius of ground essv
. a
di.""'ri
for Receiver position R ti , reflected multi-pole point
(Xr,.uIt Yrm uIt z rmult was derived by satellite position -oiopplershift
instanitaneous position of
-
istY
with satellite's elevation angle and RHCP position. moving ater srface
H
+H tl
(HlRCPLH p lumue Compute compute:
n, oiF soil miure
i-n cote
v
aground 2 Velocity of stream o -Soi
now 7 moisture -S;oil contenit%
2 05
Ro ri g )2(g
2
) +(Zg )2
2 e-' dielectric of Soil
moisture
Sand & % Clay
by -.(Sand,%(Clv)
R= {(X, - xis + (y y s )2 + (z _
z is )2 }O.5 Figure. 4. The processing diagram of reflectivity and Ground object detection
of simulation program
LXrmuit j[ r ground ( ) lXs
S procedure repeats until the following condition is met as
~rmult IR d +S 1<()
xi jHLg
zs
H I < .A
° meter Analternative is to use suitable
-d -h
Zrult X i
j r
DT

relative heights. Then, the present authors use GPS data for
The value of R ti is ellipsoid Radius of Earth reference to
reflected surface° position .The value of R rmult is Radius of
0
HRF[cp and iHoCP for the subtraction as well as the Digital
Terrain Elevation Data (DTED) level 2 (-90 ft pixel) mapping
cross point between Reflected position and line of sight: Eq. With h The value of h is provided from the output
~ ~~~~~~~~~~wthho
(3)~
~ ~ RUao condvided byo Eqt2)
(3) of R,~mult was conducted by E (2) The Eq (1), Eq. (2),ofwas
au
h he
of the GPSsolution tool in the GPS receiver 2) Eq.utput.
to Note Book
and Eq. (3) conduct as following Eq. (4) Reflected point (control log command: BESTPOSITION). Hence, the ground
center (X LI YLI zL)i; EL: elevation angle of satellite or water surface altitude of Taiwan's Local Height is
described as follows by Eq. (5).
Ri - J{X R'2+
)y
1'rmu )2 +(Zi
A (Y 2 2
rmult l rmult } \rmutt +rml JVJ +(z 21J
0.5
{3A

ti F
XLl R= Xrult (4) iRHCP 2H LHCP) (5)
I i,lt
rm ground -u2multn

183
 Proceedings of the IEEE/OES/C¶ktTC Ninth Working Conference on Current f2leasurement Technofogy
The surface roughness parameter based on the ratio of surface
height standard deviation to length of reflected footprint is
{red green blue} wavelength bands {V, V2 V3,
described as following in Eq. (6) and the two GPS-derived features by Li and L2 Band; The Eq.
pfol sand vEvia. .of
descrbeds n

(7) conducts as Eq. (8) as following [3]

t tandard veviation of_gn (6 t
slength_of reflected _foot- print v red (co Irow)
Both methods yield similar results that compare well the ti {(,row)
vg (col/row) (8)
ground-based measurements of stream water level and soil { vbu
blue
ground height. The comparison with any independently 9t 2
measured elevation still suffers due to the residual multi-path With elements that vary, depending upon the instantaneous
errors for height for RHCP receiver positions. It is hoped that GPS reflected footprint points, which were, collocated image
removal of the multi-path errors will allow comparison with pixels by using the MATLAB image tool function.
ground-based measurements to yield accurate reflected points The factor F(m, Dis<%&lnd %C7a) of soil roughness, and
and stream water level estimations. v s
volumetric water of soil moisture effect correction [13] and
soil content function [14] can be applied into the F(Disps)of
C. Reflectivity and Ground Object Detection stream and soil surface roughness effect correction and
The reflection or scattering of the GPS signal by an roughness of stream surface function is depended as following
object depends on the surface roughness and electrical Eq. (9); <S: an emissivity of stream water, C': a dielectric
properties (conductivity and permittivity) of the reflecting
material. Indeed, different materials, including dry ground constant of stream water. The iS is relational function for
(grass), wet ground (different composition of sand, silt, and 6 as following as [5]
clay), tree, forest, fresh stream water, ocean water body, road,
and concrete may provide different reflectivities. The surface 1 2
reflectivity 9j , which is defined as the ratio of the reflected s =I- _
power to the direct power, can be used for the classification of
ground objects. Dinesh Manandhar [1] has shown that the F(Dispi )=(1-t)Iexp(Displ) (9)
resultant reflection coefficient is the sum of the co-polar and D s s
cross-polar reflection coefficients. As GPS signals are When th refle point an be d r d d
circularly polarized, the reflection coefficient is indeed a
function of the elevation angle from the receiver to the ground object can be identified at each epoch, the stream flow
velocity can be roughly estimated by computing the variation
satellite.
Ticofcetsof the reflection point as a function of time and processing the
computed once the elevation angle is known. The surface Doppler measurements. The situation is depicted in Figure. 5.
reflectivity 91's is related to the reflection coefficient and the
relative signal-to-noise ratios between the reflected signal and
the direct signal. More precisely, it can be represented by the
resultant Eq. (7) is calculated for approximately specula Bare Soil
conditions as

SAR'-n
9w =F
SCs (El! S'C).SNsp5,-%Clay
% r.-F(m Disp',Sand %ly
as d
rs
(7)
where 9 Reflectivity of L1/L2 Band by Directive signal
observation or Reflected signal observation, where SNRS. and
SNR'. are the signal-to-noise ratios of the reflected and direct
signals; n~, average instrument noise of antenna's gain
amplitude and receiver's system loss of direct signals and
Reflected signals;
where F c ( EL' C ti) :A calibration scale factor by L1/L2
Band at different elevation angle of satellite environment; the
F
E~cU)fntini eulto 2 205an h Figure. 5. The Reflection point at a moving surface like as stream flow with
grazing angle depended on ground reflective materials and RHCP and LHCP receiver position for Integrated GPS receivers and measurements
satellite's elevation angle EL' Each data classification's at Wu-Chi fisherman pier
 Proceedings ofthe IEE/EOES/C¶ktTC Ninth Working Conference on Current f2leasurement Technofogy

Xti is the function of reflection signal.. The Lorentz .2 -k k 7 .2 cos(a)cos2 (8)2cos(a)cos2 (/3)
transformation A was showed as Eq. (10). The Eq. (10) has 2k
()si2(2)cos(a)si2()
2

been used to relate the wave four-vectors between the ( y2)

incoming wave and reflected wave [M0ller, 1983]. A similar kz 2,
_ sin(a sin (a
approach is adopted for the determination of stream flow.
- 2' -2' y -' where A is the wavelength. Hence, the Doppler frequency of
7vx 1+v7 2 V 'y q ' (10) the reflected signal becomes
A
A= ,
2v
c

v
2
v
v v

2 v fr 2fdv 2v
f =d-xcos(a)_cos2_()- Y_cos(ac)sin2(,8)-
2 sin(a)
___ ____x_
V2
1+7 y
V2
77y ~2 COSrujO13p (3

yvz QUrv yvv V - The two angles can also be approximated by

c Vi VY V2 sin(a) -3(i(hground)Idground [G. Beyerle, 2002],
Consider a frame centered at the reflection point; the z- cos(a) d(hRHCP- hground)Idground, sin(/5 -6(d: 0 )Idg0l,and
axis is perpendicular to the surface, the x-axis is toward the cos(/) (dground - S(dground))Idground Within a time interval, the
receiver, and the y-axis is obtained by the right-hand rule. reflection angle changes by (sin(r)) (3(a))t, csin6)) (3(A6)
and the frequency shift changes by
V
Assume that the flow velocity vector in this frame is V IV I

L v.j E J/3(Time) (fsr fd )ti according to Eq. (14), and thus

and the speed is v= v + The wave four- ty-)i (I
fsrfsd , -2[
0 2 Z1 fd I cXTimA ° ) [vrcosEz)cos (6f+vycos(Orsin2(,6i+v sin(a] (14)
(4

vector of the incoming signal is denoted by L2z/fd]k 4(Tm~ processingAdiagram

y
The
Figure.
1.0

program was shown as

of stream flow of simulation
6. The classification of ground
where fd is the Doppler frequency and c is the speed of light.
According to Snell' s law, the Doppler frequency of the Compute
reflected signal is related to the Doppler frequency of the -RHCPpositions
L
direct signal and the flow velocity by the following
formula:
0( +2__2- V _2_2 V, -2 Vz 2'72 fd c
___ y2
z f, Ic N2

-=AX t= NA NA NA NA Comput Doppler Shiftskfor each

ky, ~~NA NA NA NA kyd
~ kz, NA
L _j
NA NA NA Lkzd j Reflected point within GPS observations
l5l15Hz
=1-
f
(22 1) 2y k
-27
2y
-k 2yd 21kzd
2z f,/c -
2zfd vx 272+21+2 C k +V V 7 2 k + 72 kcompute:
c C C2c2 c 2u i f -Instantaneous vertical elevation angle relation
1here
z. v,~~~~~27
2z f . Y- v y2 k +vxvzV2y kzd
Soil Met ctModel:
|itr:Xuei ofItm;gi_v1UvCtgee0in re3ceer point
refleae point and RHCP
2-
ky, 1-2?fd 2y +v
d <y272 1 (2x
+VYV L J )kxd ++2y ~ 7 ~) k +V/ L2>7 Ikzd
yd( c ) 2|rzsa clYlslicc
The soil moisture modela |,.!;fll,>lfil
2zfd- vzn2 arit
2 + Yvvz27<kd be+ Vvz27 kd+r1+ 2Jkzd
2
"later
coastal seacontent
and of
soil su fcc
2
recesier point:Io
of reflected point and RHCP
c c c2 c2 Syd9 d/2
2 2 v ysi.~
volumetricou
~~~~~. of
C~~~~~~~~~~~~~~~~~~~~~~ompute:
angle~~~~~~~
betee2hflo y2k,frfd
~~~~~ diretio
y
)~~~~~~~~~~~~~
z C2_V2C2y-
Moitur Mode
kz
x-xi ~~~~-3
th - dimensional velocity of r ef ced point with
in
~~~~~~~~~~~~~~~~~~~~~~Instantaneous
vertical etcvation angle and
horizontal moving angle for Each LI and L2 Band
plae frro is the fiue
where Dopplerthfrequency
the tem ,
of the reflected signal and is anndsi ofvlmt' Doppler Shits
Isatnoshoiotaoigagerlt Lf~ ,

1- Let a be the elevation angle and ,6 be the 4 (Af1

angle between the flow direction and the x-axis in the x-y Rsoe
y2k~ y2k Time procedure for velocity And Doppleir shifts
plane. FromZ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
the figure, the three terms ,~ and Z Z ~~~~~~~~~~~~~~~~~~~Return
2Ak
ncn be- approximated, respecnetively_q, a
Figure 6 The processing, diagram of stream flow of simulation program
 Proceedings of the IEEE/OES/C¶IQTC Ninth Working Conference on Current !Measurement Technofogy

object condition is met as C s" = 8 The processing will the receiver would provide a float position and, computed
position will be provided by fixed ambiguity solution method.
compute by using Eq. (14) and Eq. (15). The data were collected on November 2, 2006 from (UTC
6(Af)t 2 V 246k/ time) 06:07:56 to 07:25:56 around Da-Ga River near Tai-
65(Hground) AD
i2(Time>2)
-
SAT
9(Tinw)=- 001922000
ARground 0-00019 22000 km
.1.0 Chung, Taiwan. The results of a comparison between a high-
ground ground Aground
resolution
1 * (-50
zrn
cm pixels) 1l**rrimage and reflection points
* a\ satellite

)ti 2.5 Hz/Km 1/400 Hz/m; are shown in Figure 7. In contrast, the positions of the LHCP
-f
g(A
c g Hground)
- =
2415

This Case measurement =(ft

ground
3.0-8.OHzIkm
I(Xt) ( gro7md )2 longitd24,174
The velocity of the stream flow is obtained from the
wavelength of the LI or L2 band frequency, the Doppler shifts 24 1355

5(Af)t/1000Hzkm1of1fthe reflected and direct signals,

the reflection angle, and the relative distance and altitude of 24.173
the reflection points. The Eqs. (13), and (14) give rise to Eq.
DtQ
(15). The reflection angle ix changes--1085
)ti A 2/36 10B 2.3 1087 2&
by ground/ae t r(Hgrond wt 24a1725

with the distance Di round between the reflection point and

the RHCP receiver position. The stream flow vector angle 24
oni2 4341.26
cantgersambiguit
changes by ~t'rou)
by .(D
MAT
whenn . By te w,
ground) B t
wnD 24 1715
depend on the satellite and reflection points.
coAmptrlgro ana
'11und
,~~
~(AfV'

)2
DtP 1208355 120.836 120.8365 120.837
longitude (degree)
1208375 120.838

0 "
m Ppgro) rDgrod UTC Time 2006a11 02r0556o-2006 11a03 05 56
VI."
ref)i D4JJgrod
(D1' )2
gr (15) (GPS BESTPOSITION DUlation Height:-18.818 m
Figure. 7 The reflected area and receiver positions
1r vl (0H _cpH gwd
~Dgrowl)
D2 with altitude estimation on Lon Ann Bridge

~2 ¼(f"~romd
100 roHgrov)
~ and RHCP antennas are LHCP fixed position (N24.1721 1,
E120.83730, 419.282 meters) and RHCP fixed position
(N24.17182, E120.83732, 446.682 meters). As for the
ASUREMENTSANDRESUreflection
correctiI.ons
III. MEASUREMENTShAND RESULTS
points, signals 8, 11, 27,
are reflected from the stream water and stream soil moisture
the satellite PRNs and 28

Giving 4 or more pseudo ranges and ephemeredes from and the satellite signal PRN 19 is reflected from the ground.
receiver modified an Altitude loop by least squares searching The altitudes of the reflection points on the stream are 434.126
for receiver position with accuracy altitude condition. m, 434.083 m, 434.147 m, and 434.160 m. The comparison
Modified integer ambiguity method for fixed position with analyses of stream water level, RHCP altitude, and LHCP
RHCP and LHCP antenna shall be quickly computing the
integer ambiguity by MATLAB tools. By the way, A
altitude without correction of 18.8 meters were hi
comparison analysis concerning Stephen and zuffada [15] shown as Figure. 8. The altitude of the reflection point on the
reported "5-cm- Precision aircraft ocean altimetry using GPS
reflections" and Kashani [16] reported "Once the ambiguities45.....
are correctly resolved, centimeter-level instantaneous RTK 440 R @1 2
positioning is assured, Slight mis-modeling of troposphere IE4r
corrections affects the height component quality" by using 430
instantaneous ambiguities resolution in Long-Range RTK. 42-

The insight gained in developing -5-cm altimetry at ground 4

altitudes, combined with the quantitative ground based 200 220 240 260 280 300 320 340 360
measurement and methodology. reflecdLP tied
d pnts
A. Determination ofReflection Points

altitudeuwithout correctionpofnh__ 5&8183ern

 Proceedings of the IUE/EOES/C¶ktTC NAintIh Working Conference on Current f2Vleasurement Tfeclhno fogy
ground for PRN 19 is 455.195 m. Because the instantaneous Authors think the results for reflected GPS signals and
fixed ambiguities and atmosphere correction process were reflectivity from stream flow with smooth surface
essentials to eliminate some useless observations and- The 9~ 0.22- 0.59 reflectivity of
measurements. The mean values of residuals refer to useful Dsp~-.1 .4
and reasonable for accuracy of altitude. The length of reflected stream flow measurement for PRN 8 is higher than the
footprint for PRN 28 is 480.5 meters and the standard 9~ 127 0.09- 0.185 reflectivity of PRN 27. The reflected angle
deviation of altitude AH' for PRN 28 is 5.52 meters. E'=4~o R sntsmlra elce nl
The surface roughness of stream flow and soil moisture for
PRN 28 reflected point Disp'28 5.52 -.1
E'
E27
of PRN 27. Author thinks the ns: average 60O'
=-.0 is almost instrument noise of antenna's gain amplitude and receiver's
480.5
like as smooth surface. The length of reflected footprint for system loss of reflected signals could be effective to the
PRN 27 is 115.2 meters and the standard deviation of altitude reflectivity of stream flow and soil moisture measurements.
AH'27 ground for PRN 27 is 5.21 meters. The surface roughness The reflectivity of stream flow is between 0.09 and 0.59 for
of stream flow for PRN 27 reflected point stream water body at GPS LI Band. The reflectivity of stream
Dip7= 5.21 -0 4 is almost like as smooth surface. soil moisture is between 0.02 and 0.18 for stream soil moisture
1ip7 I15.2 =045at GPS LI Band.
B. Classification of Object Detection and Reflectivity Analysis C. Stream flow and Mean velocity Analysis
The Matlab software display and remote sensing of the The comparison of stream flow and stream soil moisture
results of the reflected surface area for stream soil by using Doppler shifts, stream surface reflected points
moisture[17]. The system with integrated receiver and Matlab instantaneous rate of altitude H and range Dground
software in custom's request environment can be used to ground
detect the elevation of water body and ground surface through conduct the mean velocity of stream flow are predicted as
out the reflected GPS signals. among 1.132- 1.285 m/s (for PRN 8) (show as Figure. 10.);
Author shows GPS L1/L2 Band reflectivity 9iV 9V' with
stream flow surface at Lon Ann Bridge with Three satellites
(PRN 8, PRN 27, and PRN 28). Comparison of variation 2A
reflectivity versus stream flow, stream soil moisture, surface:20-
roughness, and reflection coefficient method at GPS L1/L2 -0 . pe R

Band frequency was shown Figure.

as 9. 2 290 3C0 310 320 030 380 50 80 290 00 3 320 330 D2Pp3P~~~~~~~~~~~20-
DO30031D 320 33 34JO35 20 30 30 20 30 opl-, NR11
W 8 1 So-w! +
9-8,-v 18 008 q05.0108~ Doppl3mPRN 27
Comparison of reflectivity for ground characterictics and elevation angle and reflectivit for L/L2 band2

Prop gation angle P op g n angle V127 4~~~~~~~J0O 0C05 0 001 8 000

nMV 111
W~vBt 170

2 e C.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~1an0(o~
2 C. V27 2220 2V2 oDd
20 90 ~~~~~~~~~~0
28~~ 29 00 1t 320 30 30 W2~ 9 0 2 3 4 5

C 60 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~12 02 V127x

-~~~~ 04 ~ ~ ~ ~ ~ ~ ~ ~ Stea fo ~ ~ ~ ~ ~ ~~~~~~~~R1tg 8 111~~~~~~~~~~~~~~~~~~~L~~5rem-

I(1~~~~~~~~~~~~~~~~J1flbUlI
Rltgl9 2~ ~ ~ ~ ~ ~ 0 280 300 ~~~~~~~~~~~310 30 333 340 35 M 290 300 30 340350320
33D

02 ~~~~~~~~~~~~~~~~~~~~~~~Rltg 27
Iw4~~~~~~ ~~~~~~~~Rltg
..... 28 UTC Time 2006 1 0206 52 30 -2006 11 02 06 57 57 s

~~~~~~~~~~~----------;~Strea fic
__04 Figure 10. Comparison analyses for stream flow Vti and V1i for PRN 8, PRN
0.2
~~~~~~~~~~~~~~~~~27,
lz 2z
0.2
and PRN 28 at Lon Ann Bridge, Tai Chung
Tie20102 06 47 560 2006 1102 0657 60(fr n
(1613 370040 370676){11 02 14 47 -1102 14 57LT) 0.277-0.409 27180438I/
nZ (o R
ni/s (fr R
28). The mean velocity for PRN 28 has zero velocities result
Figure 9 The reflectivity 9¶ 9~2 elevation
1 2 angle EL's and reflection
of stream soil moisture with timetable than mean velocities for
~~~~~~~~PRN
27 and PRN 8. The stream flow, stream soil moisture,
coefficient F :for reflected point of PRN 8, PRN 27, and PRN 28 at Lon AnnanDopeshfswtechrlcedrjcoypits ee
Bridge
The reflectivity for PRN 28 that is not at similar
and described by Matlab Tool. ~~~~~~~~~~~~~resembled
EL2828
= 19.5' elevation angle as PRN 27
EL27 = 60O, but at IV. SUMMARY AND CONCLUSION
different surface roughness on reflected ground object The results demonstrate the usefulness of the GPS
 Proceedings ofthe IEE/EOES/C¶ktTC Ninth Working Conference on Current f2leasurement Technofogy
applicable to scale of flood flow and, thus, the measurements
could be used for routine stream flow and stream water level " An Evaluation ofON
measurements as well as to detect peak of stream flood.
measuremes
This
wl d. TTRANSACTIONS
38. NO. 6. NOVEMBER 2000.
as as
paper has presented the results of reflectivity simulations
performed. It has shown that the accuracy performance of the
developed algorithms exhibits robustness with respect to the
GPS L1/L2 signature distributions for stream flow, stream soil
moisture and the fluctuation of reflection on stream flow
surface. This algorithm could be used for the recovery of
reflectance spectra from stream flow measurements and
remote sensing of flood. In the conclusion, It was found that
the classification accuracies was improved about 55-65 % by
using the multi-spectral (green, red, blue and reflectivity) data
coupled with roughness surface parameters. It is hoped that
improvements in GNSS - remote sensing resolution ability
and results of further research employing stream soil moisture
models of land-cover. The Authors issue that is the subject for
future research is the optional choice of flood or flood plain
boundary. This paper can guarantee optimum performance b
in
nf p
estimating stream classification and stream water level.
ACKNOWLEDGMENT
The authors would like to thank Mr. Chen Houng-Shenq
chief manager of SOKKIA's vendor in Taiwan for supporting
measurements by two receivers and three LHCP antennas. The
author also thanks Dr Wickert for his advise and assistance
with the instantaneous troposphere delay effects iS concerning
withereal
thme
the real instantaneous
tme trandpheirresdelay
weather data a fcthisehon
their researching urning
method during
his visit to NCKU, Taiwan. The author also thanks Dr Bud
manager of Lockheed Martin/Fort-Worth for his advised
discussion with content of oral paper.
REFERENCES
[1] Dinesh Manandhar, Ryosuke Shibasaki, Hideyuki Torioto " Prototype
Software-based receiver for Remote Sensing Using Reflected GPS
Signals " ION Proceedings, GNSS 2005.
[2] Jan P. Weiss, Penina Axelrad, Steve Anderson " Assessment of Digital
Terrain Models for Multipath Prediction at Geodetic GNSS Installations",
ION proceeding, GNSS 2006
[3] Michael S. Grant, Scott T. Acton and Stephen J. Katzberg "Terrain
Moisture Classification Using GPS Surface-Reflected Signals", IEEE
GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 4, NO. 1
JANUARY 2007.
[4] Yutaka Kobayashi, Kamal Sarabandi, Leland Pierce, M. Craig Dobson
JPLGEOSCIENCE
TOPSAR for AND
Extracting Tree Heights " IEEE
to
REMOTE SENSING, VOL.
[5] F.T. Ulaby, R.K. Moore, and A.K. Fung, Microwave Remote Sensing,
Active and Passive, vol. 1. Norwood, MA: Artech House, 1987
[6] Zuffada C., Zavorotny V., " GPS reflections for Oceanography, ice and
surface hydrology", UCAR Summer COSMIC colloquium, Boulder, CO
[7] Shen, L.C., Juang, J.C, Tseng, C.L, Tsai, C.L., Chang, C.C, "New
application of Reflected GPS Signals L1/L2 observation Techniques with
an Integrated GPS Receiver for Remote measurements and Digital
Terrain Elevation Mapping" IEEE Instrumentation and Measurement
Technology Conference IMTC 2007 1-3 May 2007 Page(s):l - 6 10.1109
/IMTC.2007.379233
[8] Shen, L.C., Juang, J.C, Tseng, C.L, Tsai, C.L., "Remote Sensing Stream
Flow Discharge & Soil Classification by Using Reflected GPS
Observations, LI & L2 Reflectivity and Digital Terrain Model" 2007/09
[9] Lowe, S. T., Zuffada C., Chao Y., Kroger P., LaBrecque J.L., Young L.E.,
"Five- cm - Precise ion Aircraft Ocean Altimetry Using GPS
Reflections," Geophysics. Res. Let. Vol. 29, No. 10, May 2002.
[10] Clyde C. Goad and Kai Borre RECPOS Least-squares searching for
receiver position for MATLAB. MATLAB software supply. 19-april-
1996.
[11] Peter Joosten Integer ambiguity estimation using LAMBDA (basic
version) for MATLAB. MATLAB software supply. 19-MAY-1999.
Mathematical Geodesy and Positioning Delft University of Technology.
[12] Teunissen, P.J.G., De Jonge P.J., and Tiberius C.C.J.M. The least-
squares ambiguity decorrelation adjustment: its performance on short
GPS baselines and short observation spans, Journal of Geodesy, Vol. 71,
No. 10, pp. 589-602. (1997)
[13] Clari, V. K. Chris L., Stuart 0. "Capacitance Sensor for Nondestructive
Measurement of Moisture Content in Nuts and Grain" IEEE Transactions
on Instrumentation and Measurement, Vol. 56, NO. 5, October 2007
[14] Chang, Z.Y., Blagoy, P. Frank J., Gerard, C., "Extending the Limits of a
Capacitive Soil-Water-Content Measurement" IEEE Transactions on
Instrumentation and Measurement, Vol. 56, NO. 6, December 2007
[15] Stephen T. Lowe, Cinzia Zuffada, Yi Chao, Peter Kroger and Larry E.
Young " 5-cm-Precision aircraft ocean altimetry using GPS reflections"
GEOPHSICAL RESARCH LETTERS, VOL.29, NO. 10, 10.1029/
2002GL 014759, 2002
[16] Israel Kashani, Pawel Wielgosz, Dorota Grejner-Brzezinska " The
Double Difference Effect of Ionospheric Correction Latency on
Instantaneous Ambiguity Resolution in Long-Range RTK" ION GNSS
17th International Technical Meeting of the Satellite Division, 21-24 Sept.
2004, Long Beach, CA
[17]Shen, L.C., Juang, J.C, Tseng, C.L, Tsai, C.L., "Remote Sensing Soil
Moisture & Classification by using Reflected GPS Observations, LI &
L2 Reflectivity" 1st Colloquium Scientific and Fundamental Aspects of
the Galileo Programme, ESC 2007, Toulouse, France
188