Professional Documents
Culture Documents
3, AUGUST 1980
317
1) Propagation loss between isotropic antennasis treated. 2) Quasi-smooth terrain, not irregular,is treated. 3) The urbanarea propagation loss is presented as the standard formula.Forotherareas,acorrectionequation to the standard is prepared.
11. PROPAGATION LOSS AND EMIPRICAL FORMULA
I. INTRODUCTION
N SYSTEM PLANNINGforlandmobileradio service or service quality evaluation, it is indispensable to determine thepropagationcharacteristics. By using manyexperimental results and by statistical data processing, many authors have developed nomograms and charts to permit the calculation of expected field strengths from a given transmitter at chosen receiver locations [ 11 -[7] and made it clear that the propagation loss shows logarithmic behavior to the distance. Of all the studies, Okumuras report [ l ] is very practical, because it carefully arranges field strength and service area. Not only is the report usedas comparisondatawithotherauthorsreports [ 8 ] - [ l o ] ,b u t also the propagation prediction methods in the report have become standard for planning in todays land mobile systems i Japan [ 1 11 , [ 121 . n In Okumuraspredictionmethod,prediction curves for a basic median field strength were given with these parameters: base stationeffectiveantenna height hb frequency f,,and vehicular station antenna height h,. Fig. 1 shows one of these curves. When malung a system plan using this method, therefore, it is necessary to select the curves according to f,, and hb, h,. It is more cumbersome to use these curves, represented by 1 kW effectiveradiatedpower(ERP)perdipole,sinceit is necessary to convert the value to the transmitters power. Under these circumstances, even though the method is practical in theactualwork stage such as radio zone estimation, it is awkward to use directly in the system planning stage required to fit the parameters best. In this report, to make computational use ofOkumuras prediction methods in system planning, an empirical formula forpropagationloss,whichcorresponds to freespace loss often usedin UHF futedradio communication system planning, is derived from the prediction curves. In order to avoid complication, the following points were considered in its formulation.
Manuscript received May 7, 1979; revised January 29, 1980. The author is with the Mobile Communication Equipment Section, ElectricalCommunicationLaboratories,NipponTelegraphandTelephone Public Corporation, 1-2356,Take, Yokisuka-shi, Kanagawa-ken, 238-03, Japan. Telephone 0468-59-2794.
(dBm) = P, (dBm/m2)
+ 10 logl ( A e f f )
(1)
where
A,R = A2/4n,
A: wavelength (m).
(dB) = P, - Pr
= P, (dBW) - E (dBpV/m) - 10 log1 0 (x2/4n)
+ 145.8.
(2)
B. Okumuras PredictionCurves and Propagation Loss As Okumuras prediction curves give the received field strength at 1 kW ERP/dipole, it is necessary to transform the unit ERP/dipole from to EIRP. This transformation is accomplished by adding the difference value for power gain between the isotropic antenna and the dipole antenna. Since the absolute power gain of the dipole antenna is 2.2 dB, then P, (dBW EIRP) = P, (dBW ERP/dipole)
+ 2.2 (dB).(3)
When P, is 1 kW (ElW/dipole), therefore, P, (dBW EIRP) is 32.2 dB. Using (2) and (3), the progapation loss L , (dB) between the isotropicantennas is given bythepredictioncurvesandthe following equation:
(4)
318
110
100
lc (MHz)
h, (m)
1 1
150
900 450
~
1
'
~
12 30 4 . 5 1 1 7 . 0 1 0 5 . 5
5 0 1 2 2 .1 1 4 .1 0 3 . 0 5 0
I
i
112.0
110.0
70 100
120.5 118.0
150 1 0 8 . 0 200
106.0
114.5
121 .o
2 0,
TABLE I1 VALUE O F B
20
fc ( M H ~
1:
10
30
0
-10 I
50
1 1 1
35.0
35.0
1
~
,
~
+Log. Scale
P Scale Linear
Distance (km)
30.4
30.4 29.9
I
29.4
30.9 2 2 99 9 9. .
two regulations: 1) at each frequency f (MHz), A decreases , two by two against logarithmically increasing hb (m), and 2) when f,becomes n times as large for fixed hb, A increases in proportion to log n. Considering these points, A can be shown as in Fig. 2. From this figure, A can be presented by
A = a - 13.82 loglo h ,
- a(h,)
L, (dB)
+ B log10 R
(8)
B=--p
where a(h,) is the correction factor for the vehicular station antenna height h , (m). In the basic curves, h, is 1.5 m, and correction curves for the other heightsare provided. Therefore, it is convenient to take a = 0 dB for h , = 1.5 m and to introduce the correction equation for the other heights. 1 ) Introduction of the Empirical Formula: Using (5) and (6), A is given by the value of the field strength E (dB pV/m) at R = 1 (km), and B is determinedbytheslope of the field where strength curve. Tables I and I1 show values for A and B taken f,: from the basic median field strength curves. Table I contains
the maximum fluctuation width is about *OS, and it becomes the linear approximation error about B. Substituting (9) and (10) into (6), the standardformulaforpropagation loss is obtained by
150-1500 MHz,
319
Fig. 2.
36
Introduction of factor A .
35
34
31
30
29
50
1c 30
20
70
100
200 hh ( m )
Fig. 3.
Introduction of factor B .
and a@,) is the correction factor for h,, and a = 0 dB for h , = 1.5 m. 2) Determination of Correction Factor a(h,): In the prediction method, correciton curves for h , are given by Fig. 4. As the propagation loss L , of (1 1) has taken h, = 1.5 m Namely, the correction is presented as relative height gain to into standard, the correction factor a(h,) satisfies the condia standard Of h I = W in an urbanarea Over quasi-smoothtionthat = 0 dB for h , = 1.5 m, Ifit is assumed thatthecorterrain. rection curves which satisfy this condition can be presented by 1 ) Correction factors in a medium-small city: In the correctioncurvesforamedium-smallcity, if the horizontal axis is a , .s = tuc) 11, - q ( f c ) , (1 2 )
translated into linear it scale, can be expected that these curves will be shown by straight lines. From this viewpoint correction curves arerewritten as shown bytheplottedpoints in Fig. 5 . Since the empirical formula should be as simple and accurate as possible for usability, the approximate lines in this figureare used.
320
20
2000'
Urban area
1-
700 15 400'
200
4b
.3
! j
C
100'
10
100
3
5
,x
'
-"
a
f e
0
200
400
'ti
loo0
'
$
0
-5
7
(m)
10
VehicularStation AntennaHeighth
Fig. 4. Prediction curves for vehicular antenna height gain in an urban area.
1500
900
150
Vehicular Station
Antenna Height
h,
(m)
using
l-lOm, 150-1500MHz.
The error to the linear approximation inFig. 5 is in proporto the frequency, and is about1 .O dB when f , = 1500 MHz. As the correction curves in Fig. 4 have irregular characSubstituting (13) into (12), the correction factor a@,) is ob- teristics in h , = 4-5 m, the approximation error of this part tained for vehicular antenna heights in a medium-small city: in Fig. 5 becomes larger than the other part. Therefore, it can be estimated that the maximum error will arise in f c = 1500 a@,) = (1.1 loglo f , - 0.7) h , - (1 -56 loglo f - 0.8) MHz and h , = 4-5 m. , 2) Correction factors in a lave city: Thecorrection (14) curves are given by dashed lines in Fig. 4. These curves can be (13)tion
321
4.0
2 .o
Frequency IC (MHz)
Fig. 6.
200
MHz
IC
400 MHz
1.5
9 1 0
(m)
Fig. 7.
considered as parabolas. The following equations are approxi- transform (1 5 ) to an equation whichsatisfies the condition that a = 0 dB for h, = 1.5 m, then the correction factor a(h,) for mate expressions of these curves: vehicular antenna heights becomes ~ 3 = 8.29 * (loglo 1.54 A,) - 3.69 (dB), a(h,) = 8.29 (loglo 1.54 h,) - 1.10 (dB), f,< 200 MHz
(1 6)
As shown in Fig. 8, values given by (16) and values given by Fig. 4 agree well in f,2 400 MHz. Their maximum difference
322
plots : Fig. 5
f 5 2 0 0 MHz
,
1.5
2
9 1 0
V e h i c u l a r S t a t i o n A n t e n n a H e l g h t h m (m)
= 900 XIHz
a
L?
180
170
160 150 140 130
%
4
1
i
d
180
170
?
i
a.
$
m
&
m
0
120 110
1 00
110
90
80
90
80
o
, 8
o
1
0 ,
~
I
30
40 50
io
60 1
70
60
10
20
30
40 50
20 7
10
Distance R (krn)
Distance R (km)
Fig. 9.
occurs in f, < 200 MHz and h , 2 5 m, and is about 1 dB. Therefore, (16) is used as the correction factor a(h,) for vehicularantennaheightsina large city, where the building height average is more than 15 m.
111. ESTIMATION OF THEAPPROXIMATION ERR-OR
In this section we investigate how accurately the empirical formulasexpresstheprediction curves. Figs. 9 and 10 show the propagation loss in an urban area with parameters f, and h b , respectively.The solidlinesare the values given bythe
formula and the dashedlines are the values given by the prediction curves. Fig. 9 shows that the error at both ends of the frequency range is very small, and the maximum error, which occurs in the middle frequency range, is only about 1 dB. Furthermore, it shows that the error is independent of the distance (1-20 km) and is constant for each frequency. This is the reason that only term A in (7) depends on frequency. Therefore, the error in Fig. 9 mainly indicates the approximation error in (19) which gives termA. Fig. 10 shows that the error fluctuates about h b , and its maximum value is about 1 dB. Thisis due to the
323
900 450
5
u
L
i :
150
V e h i c u l a r S t a t i o n A n t e n n a H e i g h t h rn (m)
linear approximation in Fig. 3. From Fig. 10, it can besaid that (lo), which gives term B , is a fairly accurate equation. Fig. 1 1 shows the correction factor a(h,) for vehicularheights in a medium-smallcity.The solid lines are values calculated by using (14)andtheplottedpointsarethecorrection values given by Fig. 4. As mentioned i Section 11-C, the linear n approximation error is porportional to the frequency and the antenna height. The maximum error arises in f, = 1500 MHz and h , = 4-5 m, and its value is about 1.5 dB. Below f, = 900 MHz, the error is only 0.5 dB (except when h , = 4-5 m), or 1 dB (in h , = 4-5 m). Consideringequation simplicity, therefore, it can be said that (14) is a simple and wellapproximated equation. The correction factor a(h,) in a large city is mentioned in the previous section.When (16), there is using little error in f, 2 400 MHz, and the maximum error is about 1 dB in f, < 200 MHz and h , 2 5 m . In practice, there are few cases in whch the vehicular antenna height is above 5 m. Therefore, it can be considered that (16) is very practicable.
3 z
L
2 i <
d a s h e d line : report
10
;
2
1 .
4
0
L
0
i
U
i
6
0
Fi
"
100
200
300
500 700
1000
2000
F r e q u e n c r f c (MHz)
area.
culated by
+ (17) 5.4,
f, : MHz.
On the other hand, the open area correction factorQ, (dB) is given by the dashed line in Fig. 13. In this figure the solid line is the approximated curve given by
Qr
+ 40.94,
(19)
3 24
Frequency IC ( M H ~ )
log
fc
- 13.82
log
hb - a ( h m ) *
= (1.1
. log
'
fc
- 0 . 7 ) , hm - ( 1 . 5 6 - 1 .1
- 4.97
log
IC
- 0.8)
l a r g e city
a ih,)
=
8.29
; fc ~ 2 0 0 MHz ;
= 3.2
. (log
area)
rcz
400
M H ~
Suburban L Area
= {Urban ~p
-2
{loglo ( r c / 2 8 ) }
- 5.4
(dB)
here
fc
: frequency
hb
: base stallon e f f e c t w e
antenna helght (mj
-_-_-_30-200
(m)
Fig. 13 .shows that (19) gives a good approximated value, and taken down to one decimal place, the error becomes very large. therefore, the propagation loss in open area L,, (dB) will be Therefore, coefficients are futed down to two decimal places. calculated by V. SUMMARY (20) L p o (dB) = L p {eq (1 1)) Qr. In order to put Okumura's prediction methods to computa-
In (17) and (19), if the coefficients are taken down to three tional use in system planning, prediction curves are formularin Table places of decimals, the error becomes slightly smaller, but the izedas propagation loss.Theseresultsarearranged 111. Since the propagation loss can be treated as a formula, it equationbecomesmorecomplicated.If the coefficientsare
325
propagation at 800 MHz in the Philadelphia area, IE Tram. Veh. becomes possible to put the formula into various calculations c l . , vol. VT-21, May about system planning. However,since the formula can only be [6] T e P.hBarsis. Determination1972. A. of service area for VHF/UHF land mobile applied in restrictedranges,it is necessary to takenotice of and broadcast operations over irregular tenain, I& Trans.Veh. Technol.. vol. VT-22. May 1973. its applicable ranges and units.
[7] J . Durkin, Computer pndiction of service area for VHF and UHF land mobile radio services, I&& Trans. Veh. Technol., vol. VT-26, Nov.
1977. [8] G. D. O, Vehicle location in cellular mobileradio systems, IEE t t
ACKNOWLEDGMENT Theauthorthanks Mr. K. Izumi and K. Hirade,Mobile Communication Equipment Section, Yokosuka Electrical Communication Laboratories, for their useful suggestions.
REFERENCES
Y. Okumura et ai., Field strength and i t s variability in UHF and VHF land-mobile radio service, Rev. Elec. Commun. Lab.. vol. 16, 1968. CCIR, Recommendation 370-1, 11, Oslo, 1 9 6 6 : Report 565, Geneva, 1974. W.R. Young, Jr., Comparison of mobile radio transmission at 150, 450, 900 and 3700 MHz, Bell Syst. Tech. J . , vol. 3 I , Nov. 1952. A. P. Barsis. Radio wavepropagation over irregular tenain in the 769200 MHz frequency range, I&& Trans.Veh.Technol.. vol. VT-20, Aug. 1971. D. M. Black D. and 0. Reudink, Some characteristics of radio
Tram. Veh.Technol., vol. VT-26, Feb. 1977. [9] K. Allsebrook and J. D. Parsons, Mobile radio propagation in British cities at frequencies in the VHF and UHF bands, I& Trans. Veh. Technol., vol. VT-26, Nov. 1977. [IO] W. C. Jakes,Jr.. Microwave Mobile Communications. New Yo&: Wiley, 1974. [ I I] F. Ikegami, Mobile radio communication in Japan, /E& Trans. C ~ m m ~ r . . COM-20, 1972. VOI. [ 121 N. Yoshikawa and T. Nomura, On thedesign of a smallzoneland Veh. mobile radio system in UHF band, /E Trans. Technol., VT-25, Aug. 1976.
Masaharu Hsta (80) for a biography and photograph see page 252 of the May 1980 issue of this TRANSACTIONS.