Ann. Telecommun.
(2010) 65:385–389
DOI 10.1007/s12243-010-0161-7
Experimental investigation of mobile radio propagation
at 1.8 GHz over macrocellular dense urban regions of Delhi
M. V. S. N. Prasad & K. Ratnamala
Received: 20 October 2009 / Accepted: 19 January 2010 / Published online: 17 February 2010
# Institut TELECOM and Springer-Verlag 2010
Abstract The design of future-generation mobile communi-
cation systems depends critically on the suitability of path loss
prediction methods and their suitability to various regions. To
investigate the radio channel behavior, experimental measure-
ments in the 1.8 GHz band were conducted in the dense urban
regions of New Delhi for six GSM base stations. Path loss
exponents and break point distances have been deduced, and
observed results have been compared with various prediction
methods, and their standard deviations have been presented.
Measured regression line exhibited lowest standard deviations
followed by rooftop propagation method compared with Cost
231 Hata and Cost 231 Walfisch–Ikegami (WI) prediction
methods. Variation of observed results has been explained in
terms of vertical and horizontal propagation mechanisms
which is the novelty of this paper.
Keywords Dense urban . Path loss modeling .
Measurements . 1.8 GHz band . Prediction methods
1 Introduction
The introduction of cellular communication systems in
India using both GSM and CDMA technologies gave a big
thrust to mobile communication scenario. Recent reports
M. V. S. N. Prasad (*)
National Physical Laboratory,
Dr K S Krishnan Road,
New Delhi 110012, India
e-mail: mvsnprasad@gmail.com
K. Ratnamala
National Geophysical Research Institute,
Uppal Road,
Hyderabad 500007, India
indicate that, in India, mobile subscriber base crossed 325-
million mark ranking this country fourth with China, US,
and Russia taking the first three positions. In order to
characterize the propagation characteristics in the 1.8 GHz
band over different areas of the country, radio measurements
utilizing the existing cellular networks in urban, suburban,
and rural environments had been undertaken. The results of
this study can be applied to newly coming 3G systems going
to operate in 2 GHz band in this country since the
propagation characteristics are more or less similar in 1.8
and 2 GHz bands. Over this country, not many measurement
related radio channel studies and associated modeling has
not been carried out. Lena Melin et al. [1] conducted
measurements in macrocells in Stockholm at 900 and
1800 MHz in urban and suburban areas. They reported that
difference in path loss between the two frequencies is 6.9 dB
in urban areas whereas in suburban areas it is 9.3 dB
suggesting that diffraction loss from vegetation is substan-
tially higher for 1,800 MHz than for 900 MHz. Kurner and
Meier reported [2] a three-dimensional radio propagation
prediction model which allows mobile operators to predict
the outdoor and indoor coverage in dense urban areas. The
drawback of this study is that it requires high-resolution
accurate building data stored in raster format.
In the present study, radio measurements at 1,800 MHz
utilizing the Idea cellular network (an operator in this
region) in various dense urban areas of Delhi and
comparison of the observed values of path loss with various
models has been carried out. Path loss exponents as a
function of distance have been deduced. Also, break point
distances have been deduced, and comparison of observed
values with various models have been carried out, and their
standard deviations have been brought out. Observed
results have been explained in terms of vertical and
horizontal propagation mechanisms.
386 Ann. Telecommun. (2010) 65:385–389

2 Experimental details Gaussian variable over macrocells. L is taken from the

The carrier signals of various GSM base station transmitters
located in the dense urban/urban environments of New
Delhi area operating in the 1.8 GHz belonging to Idea
cellular network have been monitored with Nokia GSM
receiver (model 6150) generally used as a drive-in tool for
planning cellular network along with GPS receiver to know
the latitude and longitude of the mobile. The sensitivity of
receiver is −102 dBm. The base stations used in the present
study are (1) Inner circle, (2) Arunachal, (3) Indraprastha
extension (I.P. Extension), (4) Onkar nagar sector-1, (5)
Onkar nagar sector-2, and (6) Trinagar. The transmitting
powers of all the stations are +43 dBm. All the six paths are
situated in dense urban environment. The details of the base
stations are shown in Table 1.
3 Results
The observed values of path loss for all the base stations
have been deduced for distances ranging from 50 m to
3.5 km. Path loss exponents from the data have been
deduced from the Eq. 1 shown below using the observed
path losses for various distances and L0 depicting the path
loss at 1 m distance.
L ¼ L0 þ 10n log d þ S ð1Þ
Where L is the path loss deduced for various distances,
L0 is the path loss at 1 m, and d is the distance in meters, S
is the shadow fading in decibel. Using the approach given
in reference [3] of Erceg et al., the path loss exponent n has
been deduced. In the above equation, L0 =20 log (4∏ d0/λ)
where λ is the wavelength corresponding to 1,800 MHz and
d0 is taken as one meter. s is the shadow fading variation
and varies from location to another within given macrocell.
It tends to be Gaussian in a given macrocell denoting
shadow fading as lognormal. It can be expressed as s=yσ,
where y is a zero mean Gaussian variable of unit standard
deviation and σ, the standard deviation of s, is itself a
observed path loss values.
Using the above values, the path loss exponent “n” has
been deduced. In general, at distances close to the
transmitter, exponents of the order of 7 are observed, and
then, they fall steeply up to a value of 4 around 400 m and
remains steady for the remaining distances. A typical figure
of the above exponent variation for Inner circle base station
is shown in Fig. 1.
Another typical figure (Fig. 2) showing the comparison of
observed path losses with those predicted from COST 231
Walfisch–Ikegami (WI) method [4] for street widths of 20,
30, and 40 m and building heights of 15 and 18 m and
COST231 Hata method [5] is also presented. In this case,
close to transmitter, high path losses of 150 to 170 dB have
been observed. The estimated measurement root mean square
error is around 1.5 dB. With an effective isotropic radiated
power (eirp) of 61 dBm and receiver sensitivity of
−102 dBm, the measured path loss at best is 164 dB. Some
points close to the transmitter greater than this value appeared
in Fig. 2. These might have got originated from antenna side
lobes and were not taken into consideration while computing
standard deviations of the prediction methods. In essence,
points greater than 164 dB were not used while comparing
the prediction methods with observed values. This is due to
the plotting of all the available data in Fig. 2. WI method with
heights of 18 m passes through majority of dense cluster of
observed points at all distances. The same method with
heights of 15 m shows the predicted loss less than that of
18 m and passes through some of the points. As the street
width increases, path loss decreases for a given building
height. COST 231 Hata method underestimates the values by
10 to 20 dB. The variation in path loss for a given street
width of 30 m when building height changes from 15 to 18 m
is 5.5 dB. This figure also has regression line plotted with its
coefficients. Similarly for all the remaining five base stations,
the deduction of path loss exponents as a function of distance
and comparison of observed values with predicted values has
been carried out. The figures of these base stations are not
shown due to paucity of space.

Table 1 Experimental details of base stations

Name of base station Environment Type of Gain TX Beam width Beam width Height of TX
TX antenna antenna, dBi (horizontal), ° (vertical), ° antenna, m

Inner circle Dense urban 739495 18.0 65 7 22

Arunachal Dense urban 742212 17.5 65 7 32
I.P. Extension Dense urban 742212 17.5 65 7 27
Onkar nagar (sec-1) Dense urban 739495 18.0 65 7 24
Onkar nagar (sec-2) Dense urban 739495 18.0 65 7 24
Tri nagar Dense urban 739495 18.0 65 7 24

Height of the mobile antenna=1.5 m

Ann. Telecommun. (2010) 65:385–389 387

Fig. 1 Variation of observed 8.0

path loss exponent with distance
Innercircle, hb=22m
for Inner circle base station
7.0

6.0

Pathloss Exponent
5.0

4.0

3.0

2.0

1.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Distance,km

3.1 Break point distances
The break point has been deduced as the distance at which
the slope of the curve (the path loss exponent vs distance)
changes. In Fig. 1, it changes at 200 m. In the case of other
base stations, the same approach has been followed, and the
values are shown in Table 2.
4 Standard deviations of prediction methods
for all the base stations
The mean error and standard deviations of COST 231 Hata,
WI method for street widths of 20 and 30 m, and measured
regression line with respect to the observed values have been
deduced as a function of distance for the above six base
stations. This table also contains the standard deviation and
mean errors from a rooftop propagation model proposed by
Naoki Kita et al. [6] based on geometrical propagation and
validated with measurements at 2.2, 5.2, and 19.4 GHz. This
model consists of three regions: the direct wave region, the
reflected wave region, and the diffracted wave region. The
path loss in all the regions from line-of-sight to non-line-of-
sight can be estimated continuously. Authors claim that the
model is useful for radio zone design of a system with
adaptive modulation, study of interference consideration in
the short range intended for frequency sharing among
different systems and radio zone design for next generation

Fig. 2 Comparison of observed 180

path losses with COST 231 Inner circle, hb=22m
Walfish–Ikegami and COST 231
Hata methods for Inner circle 170
base station
y = 4.5837Ln(x) + 153.78
160

150
Pathloss,dB

140
obs loss
WI, b=20,hroof=15m
130
WI, b=30, hroof=15m
WI, b=40, hroof=15m
120 WI, b=20, hroof=18m
WI, b=30, hroof=18m
WI, b=40, hroof=18m
110
COST231HATA
reg line
100

90
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Distance,km
388 Ann. Telecommun. (2010) 65:385–389

Table 2 Break point distances

observed from data Base station Height of antenna (m) Observed break point (m)

Inner circle 22 200

Arunachal 32 400
IP extension 27 –
Onkarnagar sec-1 24 200
Onkarnagar sec-1 24 400
Trinagar 24 300

wireless access systems using microwave frequencies. The
details of the method are not reproduced here, and the
interested reader can refer to their paper.
In Table 3, standard deviations exhibited by measured
regression line are low ranging from 4 to 12 compared with
Cost 231 Hata, WI, and rooftop methods. The first two
methods exhibited high values ranging from 10 to 24 dB. For
Arunachal base station, all the three methods of Cost 231
Hata, WI, and rooftop exhibited more or less same value. In
the case of Inner circle base station, rooftop method exhibited
a standard deviation of 13.20 which is much less than that of
Cost 231 Hata and WI methods and equal to regression
method. The trend continued for other base stations with
rooftop model exhibiting less standard deviation than Cost
231 Hata and WI models. This is a very important fallout of
the present study. Furthermore, the average values of standard
deviation for all the base stations for all the four methods are
16.73, 17.68, 5.88, and 13.88, respectively.
5 Discussion
At distances close to the transmitter, due to high cluster of
buildings probably line-of-sight (LOS) component could be
missing. This might have led to high path losses. As the
mobile moves due to the availability of LOS component,
path losses decrease rapidly and remain flat with moderate
variations for rest of the distances. In the case of Erceg et
al. [3]who also made extensive measurements at 1.9 GHz at
New Jersey, Seattle, and Chicago, high path losses close to
the transmitter were observed and then, the loss was
decreasing linearly with distance. In the present study, the
path loss falls steeply up to 0.5 km approximately. The
nature of variation of path loss exponent in the present
study resembled to that of variation in Erceg et al.’s study.
Steep transitions of path loss occur when the base station
antenna height is close to the height of surrounding
building rooftops. Hence, the height accuracy of the base
station antenna is especially significant if large prediction
errors are to be avoided [7].
It has been observed in all path loss and path loss
exponent variations that dispersion of points decreases with
increasing distance. Standard deviation remains quite high
up to 500 m and then settles to a smaller value. In dense
urban areas, path loss exponents close to the transmitter
remains very high, variable, and falls steeply up to 500 m
or so and then varies between values of 3 and 4 depending
on the density of buildings. These higher fluctuations close
to the transmitter require higher shadow margins which is
very difficult to implement in practice. No clear correlation
of variation of path loss exponents with base station
antenna height emerged. High path losses close to the
transmitter were explained by Barbiroli [8] through hori-
zontal and vertical propagation phenomena. Near the base
station horizontal propagation dominates and at distances
far away from the base station vertical propagation becomes
significant. The relative importance of these two depends
on the base station antenna height. The higher the base
station antenna height the sooner the vertical component
becomes dominant. At distances far away from base station
transmitter, the effect of building height becomes less
significant. In all the variations of path loss exponents, the
trend of exponent falling steeply up to certain distance say
400 or 500 m and settling to a fixed value shows this effect.

Table 3 Mean error and stan-

dard deviation of prediction S. No Mean error Standard deviation
methods in dense urban regions
Base station Cost 231 WI Regre Rooftop Cost 231 WI Regre Rooftop

Inner circle 29.96 27.83 3.47 30.69 24.09 23.96 12.43 13.20
Arunachal 12.13 2.03 4.63 20.40 10.01 10.79 4.51 11.70
I.P. Extension −6.33 −5.93 2.57 3.14 14.95 15.26 11.26 11.39
Onkar nagar sec-1 7.35 6.71 −1.62 26.46 22.66 23.76 8.58 16.00
Onkar nagar sec2 7.63 6.99 1.29 21.28 17.60 18.78 6.59 13.22
Trinagar 7.09 6.49 2.72 16.96 23.46 24.58 7.25 16.06
Ann. Telecommun. (2010) 65:385–389
Milanovic et al. [9] have observed that, in dense urban
regions, measurements show a very interesting feature. Path
loss after the break point does not increase with distance.
This could be probably due to the wave-guiding effect of
city streets as well as the existence of radio wave
components reflected and diffracted on buildings reaching
the receiver antenna. The path loss exponent in the present
study showed similar trend.
Fuschini et al. [10] explained urban propagation consisting
of radial or over the rooftop propagation where the wave
propagates over the buildings in the radial direction and
diffracts over building tops, and the second one is lateral
propagation where several rays reach the receiver after
successive interactions with vertical building walls and
edges. Various models of ORT are Ikegami et al. [11],
Walfish and Bertoni [12], and Saunders et al. [13]. The first
two models simplify the environment into uniform building
space which affects the accuracy. In the Saunders method,
accuracy is achieved at the expense of analytical complexity.
Conventional models do not take street orientation with
respect to the radio link orientation and assume that base
station height is greater than the average roof height. This is
not correct in microcells.
6 Conclusions
Narrowband signal level measurements were conducted in
dense urban environments of New Delhi at 1,800 MHz for six
base stations. Observed path losses deduced from measure-
ments were compared with COST 231 Hata, COST 231
Walfish–Ikegami, and rooftop propagation methods. Close to
the transmitter, high path losses were observed and started
falling steeply up to 500 m or so. Thereafter, they tend to
remain flat. This has been exhibited by path loss exponents
also. Close to the transmitter, high path loss exponents of 6 to
7 were observed and fell steeply to 3–4 up to a distance of
500 m. This can be explained by means of horizontal and
vertical propagation mechanisms which is the novelty of the
paper. Till the break point, the wave was traveling in the
horizontal mode of propagation, and after the break point,
vertical propagation assumed a dominant role. Break points
for all the base stations were deduced, and standard deviations
of the prediction methods were deduced as a function of
distance for all the base stations. All stations exhibited high
values of deviations close to the transmitter and fell steeply
from 500 m onwards. Measured regression line exhibits
lowest standard deviations followed by rooftop propagation
method compared with other two prediction methods. The
useful fallout of this study is the reasonable standard
389
deviations observed by rooftop propagation model. If high
standard deviations close to the transmitter are not taken into
account, then the deviations observed are confined up to 10–
15 dB. WI method with 30 m street width exhibits less
deviation than that with 20 m street width. High path loss
exponents observed close to the transmitter can force the
operators to go for higher margins, and in this context, the
study assumes importance for the design of future 3G systems
in the upcoming 2 GHz band.
Acknowledgements Authors are grateful to Mr. Aswani Khilnani of
Idea cellular networks and Mr. Rajendra Singh of TRAI for their help
in conducting the measurements. The constructive comments of
anonymous reviewers are gratefully acknowledged.
References
1. Melin L, Ronnlund M, Angbratt R (1993) Radio wave propagation: a
comparison between 900 and 1800 MHz. Proceedings of the 43rd
IEEE conference on vehicular technology, Denver, 1993
2. Kurner T, Meier A (2002) Prediction of outdoor and outdoor-to-indoor
coverage in urban areas at 1.8 GHz. IEEE J Sel Areas Commun 20:3
3. Erceg V, Tjandra SY, Parkoff SR, Gupta A, Kulic B, Julius AA,
Bianchi R (1999) An empirically based path loss model for
wireless channels in suburban environments. IEEE J Sel Areas
Commun 17(7):1205–1211
4. Saunders SM “Antennas and propagation for wireless communica-
tion systems”, John Wiley& sons, Ltd, 2001
5. Erceg V, Hari KVS, Smith MS, Baum DS, Sheikh KP, Tappenden C,
Costa JM, Bushue C, Sarajedini A, Schwartz R, Branlund D, Kaitz T,
Trinkwon D. Channel models for fixed wireless applications.
Contribution IEEE P 802.16 3C-01/29r4, IEEE 802.16
6. Kita N, Yamada W, Sato A (2007) Path loss prediction model for
the over-rooftop propagation Environment of microwave band in
suburban areas. Elec Commun Jap Part I 90(1):13–24
7. Michaelides CP, Nix AR (2001) Accurate high speed urban field
strength predictions using a new hybrid statistical deterministic
modeling technique. 54th Vehicular Technology Conference,
Atlantic City, NJ, USA, 2:1088–1092, 7–11 Oct 2001
8. Barbiroli M, Carciofi C, Falciasecca G, Frullone M, Grazioso P,
Varini A (2002) A new statistical approach for urban environment
propagation modeling. IEEE Trans Veh Technology 51:1234–1241
9. Milanovic J, Rimac-Drlje S, Bejuk K (2007) Comparison of
propagation models accuracy for WIMAX on 3.5 GHz. ICECS
2007, Electronics circuits and systems, pp 111–114 14th IEEE
International Conference, Goa, India, 11–14 Dec
10. Fuschini F, Degli-Esposti V, Falciasecca G (2004) A statistical
model for over roof top propagation. IEEE Trans Antennas Propag
42(1):230–239
11. Ikegami F (1984) Thereotical prediction of mean field strength for
urban mobile radio. IEEE Trans Antennas Propag 39(3):299–302
12. Walfisch J, Bertoni HL (1988) A theoretical model of UHF
propagation in urban environments. IEEE Trans Antennas Propag
36:1788–1796
13. Saunders SR, Bonar FR (1994) Prediction of mobile radio wave
propagation over buildings of irregular heights and spacings.
IEEE Trans Antennas Propag 42:137–144