Large-Scale Empirical Model for a 2.

4 GHz
Wireless Network in an Outdoor Environment
Mariam Alhammadi, Mouza Almansoori, Isaias Tesfu, Daniel Habte, Irfan Ahmed, Senior Member, IEEE
Department of Electrical Engineering
Higher Colleges of Technology
Ruwais, United Arab Emirates
[email protected]

Abstract—This paper presents a site-specific empirical channel • Rician fading model is used in mixed line-of-sight (LOS)
model for large-scale propagation in a 2.4GHz outdoor wireless and non-line-of-sight (NLOS) environment.
network. The received signal strength measurements are taken • Nakagami faging model is a generalized model.
in the Higher Colleges of Technology, Ruwais campus. Ruwais
is a self-contained housing complex in remote desert about • Extended Saleh-Valenzuela model is used for mmWave
240Km from the Abu Dhabi. We characterize the 2.4GHz wireless communications.
channel in this high average temperature (22-50 C) and humid
(75-90%) environment. We use log-distance path-loss model. The The first step in a mobile cellular system setup is the coverage
signal strength measurements result in path-loss exponent of area estimation. Large-scale models are extensively used for
2.264 and the standard deviation of log-normal shadowing as
12.573dB. These values are higher than the free space path-loss the coverage area prediction. The radio signal path-loss de-
model due to the site-specific environment. The obtained large- pends on many elements of the radio communications system,
scale model can be used to analytically estimate the receive signal like; the transmitter power, and the antennas, principally their
strength in a similar condition without the use of signal strength height, general location, and the gain. This is accurate for
measurements. whatever frequency is used. The RF planning of a mobile
Index Terms—large-scale model, pathloss, 2.4GHz wireless
channel communication system requires an estimation of the path-loss
and the factors affecting the signal loss for a given radio path.
I. I NTRODUCTION This path-loss can often be calculated by computer simulations
using available RF planning tools. These calculations are
Wireless channel modeling is mainly the study how the
generally attempted when preparing system design or coverage
signal travels from one point to another in a wireless medium.
activities and this depends on a knowledge of the signal
To understand the wireless network, we need to understand
propagation properties. In this work we characterize the large-
the characteristic of the wireless channels. Most of the signal
scale channel model for outdoor environment using empirical
propagates while traveling due to reflection, diffraction, and
log-distance path-loss model. In general, empirical models are
scattering. All of these are reasons why the received signal
derived by taking signal strength measurements at specific site
power decrease. In general, the distance between the trans-
[2] while, the theoretical models are based on the fundamental
mitter and the receiver identify the signal strength in decibel
principles of radio wave propagations [3]. Empirical models
(dB) using some statistical model. A propagation model is a
provide the flexibility to measure the receive signal strength
set of wireless channel parameters, deterministic or random
with reduced computational complexity and increased the
variables, and mathematical expressions used to represent the
accuracy of the predictions [4].
radio frequency characteristics of a given environment [1]. The
The channel models used for the prediction of the path-loss of
propagation models are divided into two categories: large-scale
wireless signal are divided into two groups: deterministic (the-
and small-scale propagation models. Large-scale model is also
oretical) models and Empirical models. Due to the complexity
known as “path-loss” model where we predict the strength of
of deterministic models, empirical models are more commonly
received signal over a long distance (m/km). There are many
used. There are many empirical models which are designed
different numerical models for large-scale:
for a specific environment: LOS, NLOS, indoor propagation,
• Free space path-loss model
outdoor propagation, and at different frequencies. There are
• Ray-tracing model
numerous existing research work on the characterization of
• Empirical model
wireless channel models in different environments. The sim-
The small-scale models are used to characterize the rapid plest path-loss model is Friis path-loss model [5]. Two-ray
fluctuation of the received signal over a short distance (few cm model uses free space LOS ray and ground-reflected ray. Ray-
or mm) in a short period. There are many statistical channel tracing models use multiple path-loss exponents depending on
models for small-scale fading: the RF path, it could be LOS where there are no obstructions
• Rayleigh fading model is used to model wireless channel between the transmitter and receiver and/or NLOS where
in rich scattering environment. there are obstacles between the transmitter and receiver [6]–

Authorized licensed use limited to: Higher College of Technology. Downloaded on June 04,2021 at 19:14:17 UTC from IEEE Xplore. Restrictions apply.
[9]. Chibel et al. [10] presented a comparison of predicted
and measured signal strengths in the suburban area of Kuala
Lumpur. They validated the accuracy of log-normal path-
loss model. In [11], the path-loss modelling and the path-
loss exponent calculation has been carried out for the rural
outdoor 2.4GHz channel in two south Indian villages. Path-
loss model has been derived for conference room in [12]. The
closed multipath environment was correctly modeled by one-
slope with path-loss exponent in the range of 1.2 to 1.7. In
an outdoor urban environment of Cambridge, Massachusetts,
the wireless channel for 2.4GHz with 20MHz bandwidth is
modeled in [13]. A large variance was observed in the path-
loss exponent ranging from 2 to 5. In [14], wireless channel
inside the student dorms has been characterized. It has been
found that the 2.4GHz signal propagation inside the dorms
follows the Keenan-Motley model [15] with fitting parameters
being between 0.8899 − 0.9308.
In this work, we characterize the outdoor wireless channel at
2.4 GHz in Ruwais campus of Higher Colleges of Technology
using empirical large-scale model. The rest of this paper is
organized as follows. Section II describes the measurements
procedure, setup and data acquisition for the outdoor propaga- Fig. 1. Measurement tracks in the HCT college Ruwais
tion measurements. Section III presents the large-scale path-
loss channel models developed, while section IV summarizes in meters, and n is the path-loss exponent. The Friis equation
the conclusions of the paper. is applicable in the far-field of the transmitting antenna, i.e.,
II. M EASUREMENT P ROCEDURE with distance d > df , where df is the Fraunhofer distance
2
and is given by df = 2D λ , and it must satisfies df >> D and
A. Instruments and software df >> λ [5]. For the measurement data, the received signal
We used Cisco WAP4410N wireless-N access point as strength at distance d is described by
transmitter and WiFi Analyzer App for Android as receiver.  
λ
 
1
This access point supports IEEE 802.11 b/g/n standards. Pr (d)[dBm] = Pt [dBm]+20 log +n10 log +XσdB
4π d
B. Setup (2)
where XσdB is the log-normal random variable [16] with mean
The path-loss measurement were taken in the higher col-
µdB and standard deviation σdB . the path-loss is given by
leges of technology, Ruwais campus. We measured the re-
ceived signal strength along three different tracks outside the P L(d)[dB] = Pt [dBm] − Pr [dBm] (3)
Engineering building as shown in Fig. 1. Transmitter was fixed A. Power-distance Path-loss Model
at 1.5m height and readings were taken at different times of
In power-distance path-loss model, the received signal
the day. The measurements were done in the 2.4GHz band
strength is given by the Friis equation with a random variable
(channel 6) with a transmit power of 17 dBm.
X caused by shadowing,
C. Data Acquisition  2
λ
Signal strength measurements were taken through Wifi Pr = Pt Gt Gr d−n X (4)
4π
analyzer at regular increments of distance. Since the measured 1) Path-loss Exponent: In order to find the path-loss expo-
values were time varying due to the small-scale fading, we nent n, we plot the received signal in W against the distance in
took the average of measured readings. m and add the power trendline Kd−n to the measured data as
III. L ARGE -S CALE C HANNEL M ODEL AND S TATISTICAL shown in Fig. 2. The constant K depends on the antennas gain,
A NALYSIS carrier frequency, and the log-normal shadowing. This gives
the path-loss exponent n = 2.264, where as K = 8 × 10−7 .
In general, the Friis equation is used to calculate the
received signal power in free space as B. Log-distance Path-loss Model
 2 The average path-loss in dB at distance d between the
λ
Pr = Pt Gt Gr d−n (1) transmitter and receiver is given by the log-distance path-loss
4π
model
where Pt is the transmitted power in Watts, Gt and Gr are
 
d
the transmit and receive antennas’ gains, λ is the wavelength P L(d)[dB] = KdB + 10n log + XσdB (5)
d0

Authorized licensed use limited to: Higher College of Technology. Downloaded on June 04,2021 at 19:14:17 UTC from IEEE Xplore. Restrictions apply.
 Fig. 2. Power curve fitting Fig. 4. Pathloss linear curve fitting

Fig. 5. Comparison of measured and calculated Pr and P L

Fig. 3. Linear curve fitting

P L(d)calulated [dB] = Pt [dBm] − Pr(calulated) . The log-
normal random variable at distance d is given as
or
  X(d)σdB = P L(d)[dB] − P L(d)calulated [dB] (8)
d
P L(d)[dB] = P L(d0 ) + 10n log + XσdB (6) The histogram of the log-normal random variable is shown in
d0 Fig. 6 which confirms the normal distribution in shadowing.
where d0 is the close-in reference distance [4]. Similarly, the The standard deviation of X(d)σdB gives σdB = 12.573. The
average received signal power at distance d is given as mean µdB = 0 when average shadowing incorporated into
  KdB , else µdB > 0.
d0
Pr (d)[dBm] = Pr (d0 ) + 10n log − XσdB (7) IV. C ONCLUSIONS
d
This paper presents the signal strength measurement results
1) Path-loss
 Exponent: Plot of P L(d) versus log-distance to determine the path-loss exponent and the log-normal stan-
log dd0 when added with linear trendline, gives a straight- dard deviation for the large-scale wireless channel model in
line equation y1 = mx + c1 . From here, we get the path-loss 2.4GHz band. An empirical channel model has been derived
exponent as n = m/10 and the constant c1 = P L(d0 )+XσdB from the received signal strength measurements that charac-
as shown in Fig. 4. Similarly, plot of equation (7) gives the terizes the outdoor wireless channel in the hot and humid
trendline y2 = mx + c2 as shown in Fig. 3. environment of the Ruwais city. The study shows n = 2.264
2) Standard Deviation of Log-Normal Shadowing: In order and σdB = 12.573dB for the log-distance path-loss model in
to find the standard deviation of the log-normal random (6).
variable XσdB , we use the path-loss exponent evaluated from R EFERENCES
the above procedure. We calculate the received signal power
[1] A. Neskovic, N. Neskovic, and G. Paunovic, “Modern approaches in
using Friis equation (1) with n = 2.264 as Pr(calculated) . modeling of mobile radio systems propagation environment,” Commu-
Then Pr(calculated) is converted into dBm and find the nications Surveys & Tutorials, IEEE, vol. 3, no. 3, pp. 2–12, 2000.

Authorized licensed use limited to: Higher College of Technology. Downloaded on June 04,2021 at 19:14:17 UTC from IEEE Xplore. Restrictions apply.
 less network in dorms,” International Journal of Online Engineering,
vol. 12, no. 10, pp. 58–63, 2016.
[15] S. Zhu, B. Wang, W. Y. Liu, and H. Liang, “An Empirical
Indoor Propagation Model for Underground Parking Area,” Advanced
Materials Research, vol. 433-440, pp. 4865–4868, jan 2012. [Online].
Available: https://www.scientific.net/AMR.433-440.4865
[16] J. G. Proakis, Digital {C}ommunications, ser. Electrical
engineering series. McGraw-Hill, 2001. [Online]. Available:
https://books.google.ae/books?id=sbr8QwAACAAJ

Fig. 6. Log-normal shadowing in the measured received signal

[2] T. Wysocki, “Characterisation of the Indoor Radio Propagation

Channel at 2.4 GHz,” Journal of Telecommunications and
Information Technology, vol. 1, pp. 3–4, 2000. [Online]. Available:
http://www.elec.uow.edu.au/staff/wysocki/publications/J4.pdf
[3] J. Boyer, D. Falconer, and H. Yanikomeroglu, “A theoretical characteri-
zation of the multihop wireless communications channel with diversity,”
GLOBECOM’01. IEEE Global Telecommunications Conference, vol. 2,
pp. 841–845, 2001.
[4] I. Ahmed, S. Orfali, T. Khattab, and A. Mohamed, “Characterization
of the indoor-outdoor radio propagation channel at 2.4 GHz,” in 2011
IEEE GCC Conference and Exhibition, GCC 2011, 2011.
[5] T. S. Rappaport, Wireless communications : principles and practice.
Dorling Kindersley, 2009.
[6] R. Bonneau, M. Salazar-Palma, M. Wicks, and T. Sarkar, “A Survey
of Various Propagation Models for Mobile Communication,” Smart
Antennas, pp. 239–307, 2005.
[7] H. H. Xia, H. L. Bertoni, L. R. Maciel, A. Lindsay-Stewart,
R. Rowe, and L. Grindstaff, “Radio propagation measurements and
modelling for line-of-sight microcellular systems,” 1992 Proceedings
Vehicular Technology Society 42nd VTS Conference Frontiers of
Technology, vol. 41, no. 10, pp. 349–354, 1992. [Online]. Available:
http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=245381
[8] H. H. Xia, R. Rowe, A. Lindsay-Stewart, H. L. Bertoni, and L. R.
Maciel, “Microcellular Propagation Characteristics for Personal Com-
munications in Urban and Suburban Environments,” IEEE Transactions
on Vehicular Technology, vol. 43, no. 3, pp. 743–752, 1994.
[9] L.-o. Microcellular, H. H. Xia, H. L. Bertoni, L. R. Maciel, A. Lindsay-
stewart, and R. Rowe, “Radio Propagation Characteristics for Personal
Communications,” Ieee Transactions on Antennas and Propagation,
vol. 41, no. 10, pp. 1439–1447, 1993.
[10] J. Chebil, A. K. Lawas, and M. D. Rafiqul Islam, “Comparison be-
tween measured and predicted path loss for mobile communication
in Malaysia,” World Applied Sciences Journal, vol. 21, no. SPECIAL
ISSUE, pp. 123–128, 2013.
[11] V. Dasarathan, M. Muthukumar, K. N. Elankumaran, B. W. Tumey,
N. Buris, Thyagarajan, S. Raju, and V. Abhaikumar, “Outdoor channel
measurement, pathloss modelling and system simulation of 2.4 GHz
WLAN IEEE 802.11g in Indian rural environments,” in Asia-Pacific
Microwave Conference Proceedings, APMC, dec 2007, pp. 1–4.
[12] F. Heereman, W. Joseph, E. Tanghe, D. Plets, L. Verloock, and
L. Martens, “Path loss model and prediction of range, power and
throughput for 802.11n in large conference rooms,” AEU - International
Journal of Electronics and Communications, vol. 66, no. 7, pp. 561–568,
2012.
[13] K. T. Herring, J. W. Holloway, D. H. Staelin, and D. W. Bliss, “Path-
Loss characteristics of urban wireless channels,” IEEE Transactions on
Antennas and Propagation, vol. 58, no. 1, pp. 171–177, 2010.
[14] G. Hu, “Study on the signal transmission characteristics of 2.4GHz wire-

Authorized licensed use limited to: Higher College of Technology. Downloaded on June 04,2021 at 19:14:17 UTC from IEEE Xplore. Restrictions apply.