NASA Technical Reports Server (NTRS) 19940018324: A prediction model of signal degradation in LMSS for urban areas

6 Pages·0.39 MB·English

by  NASA Technical Reports Server (NTRS)

N94-22
A Prediction Model of Signal Degradation in LMSS for Urban Areas
Takashi Matsudo, Kenichi Minamisono,
Yoshio Karasawa and Takayasu Shiokawa
KDD R&D Laboratories
2-1-15 Ohara, Kamifukuoka-shi, Saitama
356, Japan
Tel: [phone]
Fax: [phone]
ABSTRACT for urban areas. The proposed model treats
shadowing effects caused by buildings
statistically and can predict the Cumulative
A prediction model of signal degradation in
Distribution Function (CDF) of signal diffraction
LMSS for urban areas is proposed. This model
losses in urban areas as a function of system
treats shadowing effects caused by buildings
statistically and can predict a Cumulative parameters such as frequency and elevation
Distribution Function (CDF) of signal diffraction angle, environmental (urban structure)
losses in urban areas as a function of system parameters such as number of building stories
and width, and average road width.
parameters such as frequency and elevation
angle, and environmental parameters such as
LMSS PROPAGATION MODEL
number of building stories and so on. In order
APPLICABLE TO URBAN AREAS
to examine the validity of the model, we
compared the percentage of locations where
diffraction losses were smaller than 6dB Basic Concept of Proposed Model
obtained by the CDF with satellite visibility
measured by a radiometer. As a result, it was In general, the following propagation
found that this proposed model is useful for phenomena should be taken into account for
estimating the feasibility of providing LMSS in propagation model in LMSS.
urban areas.
(1) Visibility of a direct wave from a satellite.
INTRODUCTION (2) Diffraction losses of a direct wave near and in
shadow regions.
Recently, proposals for a Land Mobile (3) Effects of multipath fading due to reflection
Satellite Service (LMSS) using handheld from buildings and ground.
terminals have been advanced. In urban areas,
signal degradation in LMSS is anticipated to be However, it seemed too complicated to
very large because of heavy shadowing effects consider the correlation between direct wave
caused by buildings. However, quality of power affected by shadowing and average power
service is still expected to be good even if users of reflected waves, so the model proposed in this
with handheld terminals are in urban areas. paper only takes the above items (1) and (2) into
Some propagation models for LMSS have account as the first step in the development of the
been proposed [1],[2], however, it has become propagation model.
ambiguous to define environmental parameters in Generally, signal degradation caused by a
these models. It is therefore difficult to apply the single building can be calculated with fairly good
models to areas where the urban structure is accuracy by a knife-edge diffraction model [5].
different. In our proposed model, we apply this single
In Japan, prediction methods for visibility in knife-edge diffraction model to a number of
urban areas have been developed [3],[4]. These buildings randomly distributed along a road.
methods have used urban structure statistics such The condition of buildings in an urban area is
as a Probability Density Function (PDF) of treated statistically as environmental parameters
building stories and building width as a function in our model. The parameters consist of the
of the number of building stories, but they can distribution of the number of building stories,
not estimate signal fading. building width as a function of building stories,
In this paper, we propose a new type of and average number of buildings per km along a
prediction model of signal degradation in LMSS road. These parameters incorporated into the
355
modelareeasilyobtainedfromapublicdata
base. (4)
v=h#_(-L+d_)dl ----h_d2
Theimportantcharacteorfourmodelistobe
abletorelatesignaldiffractionlossesindB
causedbybuildingswiththeenvironmental
d2=#(hs'z-ha)2+(x-cosec(,)) 2 [m] (5)
parameters.Byusingthismodel,wecan
calculatetheCDFofsignaldiffractionlossesin
urbanareas. where:
h = height of the building edge above the
Proposed Model and Calculation of CDF straight line joining the antenna position to
the satellite [m]
Table l shows environmental parameters used dl= distance from the satellite to the building
edge [m]
in this propagation model. These parameters can
be obtained easily from a public data base. d2 = distance from the antenna position to the
building edge [m]
To develop the LMSS propagation model for
_, = signal wavelength [m]
urban areas, it becomes important to decide how
to deal with the conditions of buildings. In our hs = building height per story [m/story]
q_= azimuth angle [degree].
model, we assume that a building of height z
[story] has a width W(z)[m], and that the depth i
is the same as the width. The relation between ihe edge height h and the
The PDF of building stories B(z) and the number of building stories z can be expressed by
the following equations,:
building width W(z) of height z [story] are given
in the following equations [3],[4], respectively,:
h:{(hs.z - ha)- x.cosec(q_).tan(0)} cos(0)
[ml (6)
where:
B(z) = _ (1)
t 0 ---elevation angle [degree].
0 (z<G)
As is evident from Eq.(3) and (4), a
W(z):5511.0-1.1exp(-0.1z)} [m] (2) diffraction loss J is defined as a function of the
edge height h. By using Eq.(3)-(6), we can
solve the edge height h at which a diffraction loss
where F and G represent an average building
gives J [dB]. Once we calculate h as a function
height in stories and a minimuna building height
of J, we can obtain the edge length Le [m] as in
in stories, respectively.
Fig. 1. At any position on the Le, a diffraction
loss is equal to J(h).
Figure 1 shows the model of a building used
in our propagation model. We assume that an To be exact, the edge length Le is not equal to
the shadow length L(z) as in Fig. 1, but we
antenna is located at a distance x [m] from the
assume that both lengths are equal in order to
building and that the antenna height is ha [m].
simplify the calculation of the CDF of signal
The antenna receives a satellite signal diffracted
diffraction losses. This assumption may result in
by the building edge. Further, we assume that
overestimation of the edge length when the
the shadow length L(z) exists at the antenna
azimuth angle ¢_is near 0° or 180°.
position x along a road.
At first, we will estimate the diffraction loss at Then the shadow length L(z), in which the
diffraction loss due to a building of z [story] is
the antenna position x and the shadow length
L(z) as in Fig.1. equal to J(h), is given by the following
equations,:
A diffraction loss caused by a building can be
evaluated from a knife-edge model with fairly
W(z)(1 + cot(_)) lm] (x< Hs)
good accuracy, and a diffraction loss J(v) in dB
is calculated by the following equations [5],: L(z)= W(z) + cot(*) {(hs.z-ha)cOt(0 ) sin(,)- x}
J(v)=6.9+201ogl'C(v - 0.1)2+ 1+v-0.1} [ml (Hs_<x<He)
W(z) lm] (He<x)
(v > - 0.7 for J(V) > 0) [dBl (3)
(7)
356
whereHs[m] andHe[m] arethedistances losses are larger than Jthr [dB]
betweenabuildingandshadowboundariesina (Eq.(1), (2), (8)).
shadowcastbyabuildingasshowninFig.1.
Step5: Calculate CDF; P(J>Jthr) [%]
InEq.(7),if He<x,theantennapositionx is (Eq.(9)).
locatedoutoftheshadowcastbyabuilding.
However,thediffractionlossesarecausedbythe Figure 2 shows one example of the
edgeatthetopofabuilding,soweassumethat cumulative distribution of signal fading
L(z)isequaltoW(z)inHe<x. calculated by the model presented here. A
Fromtheabovediscussioni,t isevidentthat person with a handheld L-band terminal for
tocalculatetheCDFof signaldiffractionlosses LMSS is assumed to be on a sidewalk in an
isequivalenttosumminguptheshadowlength urban area. The environmental parameters are
L(z) alongaroadonwhichdiffractionlossesare set at values for a typical urban area in Tokyo,
largerthanagiventhresholdvalueJthr[dB]. Japan, and hs is assumed to be 4 [m/story] [4].
Therefore,thetotalshadowlengthTperkm Since buildings registered in original public data
onwhichdiffractionlossesarelargerthanJ_ bases are higher than 4 stories or more (i.e.
canbeestimatedbyintegratingL(z)fromZtlatro height >16m), a minimum number of stories G
infinity asinEq.(8),: in Eq.(1) is set at 4 in the calculation. The
definition of azimuth angles Az is shown in
Fig.3. For example, L90 means 90° to the left of
the direction of travel and R90 means 90° to the
= D.B(z).L(z) dz [m] (8) right of the direction of travel.
thr
COMPARISON BETWEEN
CALCULATED VALUES AND
where:
MEASURED DATA
Zthr = number of stories of buildings which
cause a diffraction loss Jthr at an antenna
An Outline of a Field Experiment
position x [story]
D = average number of buildings per km
In order to examine the validity of this model,
shown in Tablel [/km]. This is easily
we carried out a field experiment in an urban
obtained from a public data base.
area. In the experiment, satellite visibility (Line-
Of-Sight(LOS) condition) was measured by
Finally, the CDF of signal diffraction losses
measuring sky noise temperature Tn [K] with a
J[dB] which are larger than a given threshold
radiometer as a function of azimuth and elevation
value Jthr[dB] can be obtained by Eq.(9),:
angles.
If Tn is higher than a given threshold value
P(J->Jthr) = 100. T [%]. (9) Tthr [K], we understand that the LOS condition
1000
is lost. Since the boundary of the LOS condition
gives a signal diffraction loss of 6dB, we
Here we summarize the procedure for
compared satellite visibility measured by the
estimating the CDF of signal diffraction losses.
radiometer with the percentage of locations
where diffraction losses were smaller than 6dB
Stepl: Give the diffraction loss (threshold
estimated by this model.
value) Jthr in dB.
The center frequency of the radiometer was
12GHz (bandwidth: 100MHz, time constant:
Step2: Calculate the building height Zthr
0.1 sec) and a horn antenna (half-power
[story] at which the diffraction loss is beamwidth: aboutl0 °) was used. All
Jthr [dB] (Eq.(3)-(6)).
experimental equipment was installed aboard a
van.
Step3: Estimate shadow length L [m] as a Calibration of the measurement system for the
function of height z [story] (Eq.(7)). determination of T_ toward the direction of the
shadow boundary was carried out at a site where
Step4: Estimate the total shadow length T
there was only one building. As a result of the
[m] per km over which the diffraction
357
measuremenat,noisetemperatureof 176.6K
public data base, this model can be expected to
wasassignedtoTthr•
be useful for estimating the feasibility of
Theexperimenwt ascarriedoutinoneofthe
providing LMSS in urban areas in many
mostbuilt-upareas(ShinjyukuareainTokyo)in countries.
Japan.Thevaluesoftheenvironmental In future, the effects of both direct wave
parametersarethesameastheonesshownin
degradation and multipath fading should be
Fig.2exceptforx=10m,ha=3mandR=29.1m. related to environmental parameters.
Thevanwasdrivenroundacourse,thelength
ofonecircuitbeingabout6.5kin. Theantenna ACKNOWLEGEMENTS
directionrelativetothetraveldirectionofthe
vehiclewasmaintainedconstanftortheperiodof The authors would like to thank Dr. K. Ono,
eachexperiment.Thedefinitionofazimuth Dr. Y. Takahashi and Dr.Y. Ito of KDD for their
anglesAzinthisexperimentisshowninFig.3. encouragement throughout this work. This work
was supported by INMARSAT under Contract
Results Project 21 R&D Studies(INM/92-798/BK) for
KDD.
Figure 4 shows examples of measured noise
temperatures for elevation angles E1 = 20°, 40° REFERENCES
and 60° at Az=R90. Each dashed line means a
noise temperature of Tthr. From this figure, we [1] CCIR, "Factors affecting the choice of
can clearly recognize that the visibility depends antennas for mobile stations of the land mobile-
on elevation angles. satellite service," Rep.925-1, vol.VIII,
Figure 5 shows a comparison between ITU, 1990.
calculated values expressed as lines and [2] E. Lutz et al.,"The Land Mobile Satellite
measured data expressed by symbols. For Communication Channel-Recording, Statistics
example, the line R90 means calculated values and Channel Model," IEEE Trans.Vehicle.Tech.,
for Az=R90 in Fig.3, and the symbol MR90 vol.40, no.2, pp.375-386, May 1991.
means measured data for the same Az. The [3] E. Ogawa and A. Satoh, "Path visibility and
calculated values represent the percentage of reflected wave propagation characteristics in
locations where signal diffraction losses are built-up areas," IEICE Trans., vol.J69-B, no.9,
smaller than 6dB estimated by the model pp.958-966, September 1986.
presented here. The measured visibility is [4] Y. ho, "Propagation characteristics for
defined as the percentage ratio of the distance for mobile reception of satellite broadcasting in
which Tn_<Tthr to the length of one circuit (6.5 urban areas," IEICE Trans., vol.J73-B-II, no.7,
kin). pp.328-335, July 1990.
As seen from this figure, the agreement [5] CCIR, "Propagation data for land mobile-
between calculated values and measured data is satellite systems for frequencies above
excellent in predictions for cases having 100MHz," Rep.lO09-1, vol.V, ITU, 1990
visibilities more than 50%. However, we can including the contribution document to CCIR
recognize some discrepancies between the WP5B (Dec., 1991) from Japan (Doc. 5B/3-E).
calculated values and the measured data for cases
having a measured visibility smaller than 50%.
This discrepancy rnay be caused because
buildings with a height of lower than 4 stories
(i.e. height < 16m) are completely omitted in the
calculation.
CONCLUSIONS
We proposed a new type of prediction model
of signal degradation in LMSS for urban areas.
Good agreement was confirmed between satellite
visibility estimated by this model and that
measured by experiments. Since we use
environmental parameters easily obtained from a
358
Table 1. Elwironmental parameters used in the propagation model
notation unit
parameters
building stories Z [story]
building height per story hs [m/story]
average number of buildings per km D {/km]
PDF of building stories B(z)
building width ( as a function of z ) W(z) [m]
average road width R [m]
Satellite _: Azimuth angle [degree]
O: Elevation angle [degree]
X "_ x: Position of antenna [m]
\ \ d, \ ha: Height of antenna [m]
-_-T, ,, A f._',
/ ',....i.,,,_ \ _l,_;_z)',,
_,V"-'Uvz:'\?-I / ',
Direction of travel __, ,'--Le--__/'_
Figure 1. Model of a building for estimating signal diffraction losses.
359
Az=R90 I
300
99.99 .... I .... I ........ I ....
99.9 -.._.-..+_:.++i.............. _..:_...................,..................-,................. ID,_
N g i...........L.. i 200
,,mO,_ 99 ....._....r.,_,_._ ..........._...................._...................i,--
o ............._.',k:........."...+"<= ......i.................t................... 100
O ....._.._................ _............... _............... 7,,._._.._...22.T.................
-_-- ........_...............:./......................
z 0
E_ZZZL- IO0 200 300
O 5O (a) Elevalion angleof 60°
_8.................. :.................. ;...................................... yI.................
.................. -.................. -................... _.................. *.................
10 ......._El=30°i .................i.................._...................
5
45°
ej 1
........ 6(F
.1
.................................. ;...................... i.....................................
.... i .... l .... i .... i ....
.01
-40 -30 -20 -10 0 10 lE3
0 1O0 200 500
Relative signal level [dB] (l'l) Elevatiol; angle of 40 °
Minimum Stories [story] :G = 4.000 300
d
Mean Stories [story] :F = 5.800
Average Number of _ia-J 200
t-.
Buildings per km [/kin] :D = 26.460
iO0
Azimuth Angle [degree] : qb = 90
z 0
Position of Antenna [m] : x = 2.000 0 1O0 200 300
DISTANCE (P1)
Height of Antenna [m] : ha = 1.500
RoadWidth [m] : R = 30.000 (c) Elevation angleof20°
Frequency [GHz] = 1.540
Wave Length [m] : X = 0,195 Figure 4. Examples of measured noise
temperatures for three elevation angles
(Az=R90). Each dashed line means
Figure 2. One example of cumulative Tth r (176.6K).
distribution of signal fading calculated
by the proposed model.
100
]_. R _._] Buildings
8O t ,, i/ ..............
i L............................. ............................
6o //;, ,i/-
Az=0
"_.. 4O
............7: ...............;jl .....I L90 O M[,90
_ -- '< ! > ' O - IA5 t M[A5
/ _O • MO
2O
IL45, [--- -R45 x MR45
[ - R90 a MR90
0 J
2O 30 40 50 60 70 80
Elevation angle [deg]
i
Figure 5. Comparison of visibility between
Figure 3. Definition of azimuth angles. For measured values and calculated ones
example, L90 means 90° to the left of (the percentage of locations where
the direction of travel and R90 means signal diffraction losses are smaller
to the right of the direction of travel, than 6dB).
360