![Distribution of CPE covered by two base stations. required for different environments. The simplest model is the single slope path loss model [8]:](https://www.researchgate.net/profile/E-T-Tchao/publication/259338579/figure/fig2/AS:549839412752385@1508103339891/Distribution-of-CPE-covered-by-two-base-stations-required-for-different-environments_Q320.jpg)
Content uploaded by E. T. Tchao
Author content
All content in this area was uploaded by E. T. Tchao on Oct 11, 2017
Content may be subject to copyright.
Journal of Communication and Computer 10 (2013) 863-872
On the Comparison Analysis of Two 4G-WiMAX Base
Stations in an Urban Sub-Saharan African Environment
Eric Tutu Tchao1, Kwasi Diawuo1 and Willie Ofosu2
1. Department of Electrical/Electronic Engineering, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana
2. Department of Electrical Engineering Technology, Penn State Wilkes-Barre, USA
Received: June 04, 2013 / Accepted: June 15, 2013 / Published: June 30, 2013.
Abstract: A growth in the demand for WBA (wireless broadband access) technology has been seen in Ghana in the last few years..
The reason for this growth can be attributed to the emergence of the use of multimedia applications, demands for ubiquitous
high-speed Internet connectivity, the massive growth in the wireless and mobile communications sector and the deregulation in the
telecommunications industry. WiMAX, which is a WBA technology, is currently being deployed in the 2,500-2,690 MHz band to
help serve the ever increasing needs of broadband internet subscribers in the country. This paper presents simulation results and field
trial measurements for two BS (base stations) in a newly deployed 4G-WiMAX network in a typical dense urban Sub-Saharan
African environment. We have used a model which uses the interference to noise ratio (n) parameter to obtain the coverage range of
these two BS under evaluation. The simulated cell range for site 1 is 4.90 km and that for site 2 is 3.19 km. The final coverage
simulation and the field measurements results have also been presented.
Key words: Wireless broadband access, WiMAX, field measurements, interference modeling, Sub-Saharan African environment.
1. Introduction
WiMAX is a wireless digital communications
system which is intended for wireless metropolitan
area networks. WiMAX can provide BWA
(broadband wireless access) up to 30 miles (50 km)
for fixed stations, and 3-10 miles (5-15 km) for
mobile stations [1]. WiMAX allows for more efficient
bandwidth use and is intended to offer higher data
rates over longer distances [2]. Because WiMAX
operates in both licensed and non-licensed frequency
bands, it provides a viable economic model for
wireless carriers which make it a BWA technology of
choice for deployment in many developing
countries [3].
Deployment of WiMAX networks has currently
started in many sub-Saharan African countries. In
Ghana, the 2,500-2,690 MHz band, sometimes
Corresponding author: Eric Tchao, Ph.D., student, research
fields: wireless networks. E-mail: ettchao.coe@knust.edu.gh.
referred to as 2.6 GHz band, which is one of the
various bands defined by the ITU (International
Telecommunication Union), has been auctioned for
WiMAX deployment [4].
Pilot deployments are well underway in several
parts of Accra and Tema. The first successful pilot
deployment which covers 55 km2 in the urban centers
of Accra and Tema has about 535 fixed and mobile
CPE (customer premise equipment). Eleven WiMAX
base stations have been used to provide coverage to
the CPEs in the network using an adaptive 4T4R (four
transmit four receive) MIMO (multi input multi output)
antenna configuration. The final distribution of the
antenna sites in the deployment area is shown in
Fig. 1.
In order to deploy this high grade pilot WiMAX
network, engineering tools and techniques that
allowed rapid system design were used to achieve the
main objective of planning the new network to give
ubiquitous coverage to the user terminals in the
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
864
Fig. 1 Final distribution of base stations in pilot network.
service area. This paper explores the techniques and
the accuracy of network plan which was used to
successful deploy two WiMAX base stations.
These techniques which we will discuss in this
paper were later extended to the deployment of the
extensive and high grade pilot WiMAX network.
2. Propagation Environment and
Interference Modeling
In order to estimate signal parameters accurately in
wireless communication systems, it is necessary to
estimate signal propagation characteristics in different
terrain environments, such as free space, urban,
suburb, country and indoor. To some extent, the
communication quality is influenced mainly by
applied terrain environment [5]. Propagation analysis
provides a good initial estimate of the signal
characteristics.
There are two general types of propagation
modeling: site-specific and site general. Site-specific
modeling requires detailed information on building
layout, furniture, and transceiver locations. It is
performed using ray-tracing methods. For large-scale
static environments, this approach may be viable. For
most Sub-Saharan environments however, the
knowledge of the building layout and materials is
limited and the environment itself can change. Thus,
the site-specific technique is not commonly employed.
Site general models provide gross statistical
predictions of path loss for link design and are useful
tools for performing initial design and layout of
wireless systems especially under Sub-Saharan
African conditions.
The Hata-Okumura model which is best suited for
large cell coverage can be used to model the
propagation environment for this WiMAX network.
Because of its simplicity, it is a widely used model for
most of the signal strength predictions in
macro-cellular environment [6, 7], even though its
frequency band is outside the band of WiMAX.
In order to estimate the coverage range of the two
base stations, mapping of the CPEs which will be
served by the two bases stations was done and the
distribution is as shown in Fig. 2.
The propagation environment can be modeled using
the distribution of CPEs as shown in Fig. 2. It consists
of a fixed station F00 (site 1) and a mobile station M00
attempting to establish a radio link in the presence of
n additional fixed stations
niFi1. Each fixed
station communicates with additional mobile stations.
Mi is the total number of mobile stations
communicating with the CPE Mij at a power Fij on the
downlink channel. Different propagation models are
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
865
Fig. 2 Distribution of CPE covered by two base stations.
required for different environments. The simplest
model is the single slope path loss model [8]:
o
o
tr dd
d
d
PP
,
(1)
where Pr is the power received at a distance d (relative
to the reference distance) from a transmitter radiating
at a power Pt. The parameter γ is the path loss
exponent (in free space) and κ is the free space path
loss between the transmission antenna and the
reference distance do:
2
r4
o
td
GG
(2)
where Gt and Gr are the antenna gains of the
transmitter and receiver respectively and λ is the
wavelength of the transmission. The single slope path
loss model is used to describe the mean path loss in
large area environments [9, 10]. Fig. 2 however shows
propagation at close ranges in an urban environment.
Propagation at close ranges however behaves more
like the plane earth model and a dual slope path loss
model is more appropriate [9, 11]:
(3)
where b is the breakpoint distance, γ1 is the path loss
exponent before the breakpoint and γ2 is the path loss
exponent after the breakpoint. The breakpoint is
related to the height above plane earth of the
transmitter antenna ht and receiver antenna hr and is
approximately given by [11]:
rt hh
b4
(4)
Eq. (4) gives one theoretical expression for the
breakpoint in the plane earth model, however, the
breakpoint is not well defined due to the oscillatory
nature of the signal envelope in the plane earth
model, and different definitions of where the
breakpoint occurs give slightly different expressions
[12, 13]. Over a region of tens of wavelengths, a
received signal will exhibit variation about the mean
power predicted by the path loss models of Eqs. (1)
and (3). Measurements have consistently indicated
these power variations exhibit lognormal statistics [8,
14]. This phenomenon is called lognormal
shadowing and can be incorporated into either path
loss model as a multiplicative factor to the path loss
PL:
10
10
Lr PP (5)
where ζ is a normally distributed dB variable with
db
b
d
d
d
kP
bdd
d
d
kP
P
o
t
o
o
t
r2
1
1
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
866
zero mean, and a standard deviation σ typically
between 6 and 12 dB in macrocell systems [8]. From
Eq. (5) it can be seen that shadowing models both
signal attenuation due to obstructions and signal
amplification due to waveguide effects. At a receiver,
a link will be considered successful if the signal to
noise plus interference ratio
1N
S is greater than
or equal to the system protection ratio Z, otherwise an
outage is deemed to occur. The region in which this
threshold is maintained is the region in which radio
communication is considered successful and is called
the “cell”. The extent of the cell is thus a function of
the radio signal and interference statistics [15].
Examining Fig. 1, assuming M00 is communicating
with Fo in the presence of a single interferer Fi which
spills a power Pu into the wanted uplink and Pd into
the wanted downlink. A spatial analysis of link outage
in the presence of an interferer but in the absence of
receiver noise was presented by Cook [16] using the
single slope path loss model. When the effect of
receiver noise is incorporated (assuming that the same
protection ratio applies to noise and interference),
M00’s uplink outage contour is a family of circles
centred on the fixed station F0, but the downlink
outage contour is a higher plane curve [17]. By
introducing a parameter called the interference to
noise ratio (n) the equations for both outage contours
can be written in a simple form. n is the total
interference power at a receiver divided by the
receiver noise power. For a single interferer under the
single slope path loss model, n at the fixed station F0
is given by [17]:
N
rP
niu
u
0,
(6)
whiles the DL n at the CPE Moo is given by:
N
rP
nid
d
0,
(7)
Using this parameter, it can be shown that the
equation for the uplink outage contour can be written
[17]:
1
1
1
,,
uu
u
ooiuuooo nn
n
rKr
(8)
and the DL outage contour equation can be written:
1
1
1
,,
dd
d
ooiddooo nn
n
rKr
(9)
where ZNKP
t
and the parameter is given by
[14]:
tt
sirtt
WZP
LWGGP
K (10)
where Ls is a system loss factor, Pi is the interference
power, and Wi and Wt are the bandwidths of the
interfering and wanted signals respectively.
As the receiver S = [N + I] threshold must be
exceeded on both the uplink and downlink in order for
the duplex link to be successful, the range of the
mobile terminal from in any direction is the minimum
of ru and rd. Thus, the range described by Eq. (8)
represents the cell radius regardless of the downlink
conditions.
3. Coverage Simulation
The coverage prediction for the deployed network
is based on the simulation parameters in Table 1 and a
realistic distribution of BS and CPEs in the network.
Genex-Unet has been used to simulate the average
throughput per sector and the final radio network plan.
The CPE antenna configuration used for the
coverage and capacity simulation was 1T2R.
The capacity simulation results in Table 2 and Fig.
3 show better performance by 4 × 4 Antenna
configuration. The minimum simulated cell edge
uplink rate using a DL/UL ratio of 35:12 which was
used for the final network plan shown in Fig. 4 is 256
Kbps. This value was obtained at 4.90 km and 3.19
km for sites 1 and 2 respectively as shown in Fig. 5.
4. Measurement
The measurements were done at several locations
within the network in Accra which are dense urban
and urban environments. The selection of these
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
867
Table 1 Simulation parameters.
Parameter Site 1 Site 2
Resource frequency 2.5-2.53 GHz
Channel bandwidth 10 MHz
Average users per sector 10
Fast Fourier Transform (FFT) Size 1,024
Subcarrier spacing 10.93 kHz
Useful symbol time 91.4 μs
Guard time 11.4 μs
OFDMA symbol time 102.8 μs
Modulation QPSK, 16-QAM, 64-QAM
Antenna frequency range 2.3-2.7GHz
VSWR ≤ 1.5
Input impedance 50 Ω
Gain 18 dBi
Horizontal beamwidth (3 dB) 60°
Vertical beamwidth (3 dB) 7°
Electrical downtilt 2°
CPE antenna configuration 1T2R
Maximum power (dBm) 43 43
Antenna height 42 m 38 m
Table 2 Capacity simulation results.
Permutation TDD split ratio
WiMAX carrier average throughput
per sector
Minimum simulated cell edge
throughput per sector
4T4R adaptive MIMO(Mbps) 4T4R adaptive MIMO
DL UL DL (Mbps) UL (Kbps)
PUSC with all SC 1 × 3 × 3 26:21 11.23 5.18 1.21 104
PUSC with all SC 1 × 3 × 3 29:18 13.00 4.32 1.52 108
PUSC with all SC 1 × 3 × 3 31:15 14.18 3.46 1.76 128
PUSC with all SC 1 × 3 × 3 35:12 16.55 2.59 2.24 256
Fig. 3 SINR simulation for the MIMO antenna configurations.
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
868
Fig. 4 Final radio simulation of the 2 BS.
Fig. 5 Coverage range of the 2 BS.
locations was based on various criteria, where distance,
elevation and line of sight capabilities were the main
factors. The field trial measurement setup comprised:
(1) GPS antenna;
(2) RF cable;
(3) Dongle XCAL-X;
(4) Laptop with XCAL-X software;
(5) WiMAX PCMCIA CARD.
The measurements were divided into physical
measurements and throughput measurements. The
physical measurements collected RSSI (received
signal strength indication) values in about 10,260 and
7,210 locations in sites 1 and 2, respectively. The
overview of the measurement areas around the base
station and the summary of the measured RSSI values
have been summarized in Figs. (6), (7), (8), and (9).
Throughput measurements were performed by
downloading and later uploading FTP file size of 10
MB form and to a remote server as the subscriber
moves away from the base station until the connection
was dropped. The results of the throughput
measurement have been discussed in the next section.
5. Discussion of Results
From the RSSI measurements taken from the
102,060 locations in sites 1, about 86.55% of the
measured RSSI values were greater than –80 dbm.
From the site 2 measurements taken from 7,210
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
869
Fig. 6 Summary of RSSI results for site 1.
Fig. 7 RSSI measurement areas for site 1.
Fig. 8 RSSI summary results for site 2.
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
870
Fig. 9 RSSI measurement areas for site 2.
locations within the cell, 79.39% of the measured
RSSI values were greater than –80 dbm. For site 1,
99.45% of the measured values obtained from up to a
of about 5.30 km away from the base station were
greater than or equal to the simulated cell edge RSSI
value of –90 dBm. For site 2 on the other hand,
95.81% of the measured RSSI values obtained from
up to about 3.5 km away from the base station were
greater than or equal to the simulated cell edge RSSI
value. The results of the throughput measurements
have been summarized in Table 3 and Fig. 10.
The maximum measured throughputs for sites 1 and
2 are 8.62 Mbps and 7.20 Mbps, respectively.
Comparing the field results with the Federal
Communications Commission’s broadband
applications support speed guide [18], the least
measured throughput of 108 kbps and 203 Kpbs for
sites 1 and 2 respectively is enough to support
applications like online job searching, navigating
websites, streaming radio, voice over IP calls and
standard video streaming.
Field measurement trials for the overall pilot
network which have been presented in Ref. [19]
showed that the maximum measured throughput
during the entire pilot network trial measurement was
9.62 Mbps as compared to the simulated throughput
per sector of 16.55 Mbps. The measured throughputs
of 8.62 Mbps and 7.20Mbps for sites 1 and 2 show the
two base stations overall network performance have
been very impressive when compared with results in
Ref. [19]. The two sites will undoubtedly give
ubiquitous network coverage to its CPEs given the
obtained 99.45% and 95.81% RSSI values of -90 dBm
measured for sites 1 and 2 respectively. The
propagation models used for the coverage and
capacity simulation were developed for areas where
the harshness of the Sub-Saharan African terrain has
not been considered explicitly. Analysis of field
measurement results in Ref. [20] supports the
assumption that the correction factor for
Hata-Okumara model which have not been specified
for the Sub-Saharan African environment could have
contributed to the differences between the simulated
throughput per sector and the measured values.
Table 3 Summary of throughput measurements.
Site Measured throughput (DL) Measured throughput (DL)
Maximum (Mbps) Distance (m) Minimum (Kbps) Distance (m)
Site 1 8.62 500 108 5,200
Site 2 7.20 420 203 3,500
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
871
Fig. 10 Throughput measurements results.
6. Conclusions
Simulations and field trial measurement results for
two WiMAX base stations in a newly deployed pilot
network in the urban centers of Accra have been
presented. In this paper we have made in-depth
analysis of the network parameters based on
measurements performed at locations within the
network. Analytical expressions have been used to
support the simulation methodology for deriving the
cell radius for the two WiMAX base stations based on
the stochastic distribution of the customer premise
equipment.
We have used Genex-Unet to simulate the final
radio plan and throughput per sector of an adaptive 4
× 4 MIMO antenna configuration. The correction
factor for Hata-Okumura Model will modeled in
subsequent papers and it effects on the performance of
the network discussed.
It has been seen through the measurement results
that at distances of 5.20 km, users of the network can
enjoy standard video and radio streaming services.
This goes a long way to validate the claim that
WiMAX technology can deliver last mile broadband
technology to subscribers, as have been seen, even
under the harsh conditions in Sub-Saharan African.
References
[1]. IEEE 802.16: IEEE Standard for Local and Metropolitan
Area Networks, part 16: Air interface for fixed broadband
wireless access systems, IEEE 2004.
[2]. IEEE Standard 802.16e, part 16: Air interface for fixed
and mobile broadband wireless access systems, IEEE
2006.
[3]. J.D. Sandra, C. Lee, WiMAX in Africa: the new frontier;
and Comparative and International Law Institute
Certificate Thesis, The Catholic University of America,
Columbus School of Law, May, 2008.
[4]. Selection and award procedure documentation: “Internet/
Broadband Penetration Statistics Broadband Wireless
access License in the 2.50 GHz-2.69 GHz Band in
Ghana”, Government White paper, January, 2012, 12-14.
[5]. X.M. Huang, J.J. Luo, Automatic configuration and
optimization of WiMAX networks, In WiMAX Network
Planning and Optimization, 2009, pp. 44-51.
[6]. M. Hata, Empirical formula for propagation loss in land
mobile radio services, IEEE Transactions on Vehicular
Technology, vol. VT-29:317325, Sep. 1981.
[7]. Y. Okumura, Field strength and its variability in VHF and
UHF land-mobile radio-services, Review of the Electrical
Communications Laboratory, vol. 16, Sept.-Oct. 1968.
[8]. G.C. Hess, Land-Mobile Radio System Engineering,
Artech House, Boston, 1993.
[9]. T.S. Rappaport, S.Y. Seidel, M.J. Feuerstein, K.L.
Blackard, H.H. Xia, Path loss, delay spread and outage
models as functions of antenna height for microcellular
system design, IEEE Transactions on Vehicular
Technology 43 (1994) 487.
On the Comparison Analysis of Two 4G-WiMAX Base Stations in an Urban Sub-Saharan African
Environment
872
[10]. L.R. Maciel, A. Lindsay-Stewart, H.H. Xia, H.L. Bertoni,
R. Rowe, Microcellular propagation characteristics for
personal communications in urban and suburban
environments, IEEE Transactions on Vehicular
Technology 43 (1994) 743.
[11]. L.R. Maciel, A. Lindsay-Stewart, H.H. Xia, H.L. Bertoni,
R. Rowe, Radio propagation characteristics for line-of-
sight and personal communications, IEEE Transactions
on Antennas and Propagation 41 (1993) 1439.
[12]. M. Taylor, D. Li, V. Erceg, S. Ghassemzadeh, D.L.
Schilling, Urban/suburban out-of-sight propagation
modeling, IEEE Communications Magazine 30 (1992)
56.
[13]. E. Green, Radio link design for microcellular systems,
British Telecom Technology Journal 8 (1990) 85.
[14]. W.C. Jakes, Microwave Mobile Communications, IEEE
Press, New York, 1974.
[15]. B.C. Jones, D.J. Skellern, An integrated
propagation-mobility interference model for microcell
network coverage prediction, Wireless Personal
Communications 5 (1997) 223-258.
[16]. C.E. Cook, Modeling interference effects for land mobile
and air mobile communications, IEEE Transactions on
Communications 35 (1987) 151-165.
[17]. B.C. Jones, D.J. Skellern, Interference modeling and
outage contours in cellular and microcellular networks, in:
2nd MCRC International Conference on Mobile and
Personal Communications Systems, Adelaide, Australia,
Apr. 10-11,1995.
[18]. Broadband Speed Guide, http://www.fcc.gov. (Accessed
Jan. 7, 2013)
[19]. E.T. Tchao, W.K. Ofosu, K. Diawuo, Radio planning and
field trial measurement for a 4G-WiMAX network in a
Sub-Saharan Africa environment, in: IEEE Wireless
Telecommunications Symposium, Phoenix, Arizona, Apr.
17-19, 2013.
[20]. L. Nissirat, M.A. Nisirat, M. Ismail, S. Al-Khawaldeh, A
terrain roughness correction factor for hata path loss
model at 900 MHz, Progress in Electromagnetics
Research C 22 (2011) 11-22.
