Progress in Electromagnetics Research B, Vol. 44, 191-221, 2012

Progress In Electromagnetics Research B, Vol.
44, 191221, 2012

CELLULAR SYSTEM INFORMATION CAPACITY
CHANGE AT HIGHER FREQUENCIES DUE TO
PROPAGATION LOSS AND SYSTEM PARAMETERS
K. A. Anang
*
, P. B. Rapajic, R. Wu, L. Bello, and T. I. Eneh
Wireless and Mobile Communication Research Centre, School of
Engineering, University of Greenwich, UK
AbstractIn this paper, mathematical analysis supported by
computer simulation is used to study cellular system information
capacity change due to propagation loss and system parameters (such
as path loss exponent, shadowing and antenna height) at microwave
carrier frequencies greater than 2 GHz and smaller cell size radius. An
improved co-channel interference model, which includes the second
tier co-channel interfering cells is used for the analysis. The system
performance is measured in terms of the uplink information capacity of
a time-division multiple access (TDMA) based cellular wireless system.
The analysis and simulation results show that the second tier co-
channel interfering cells become active at higher microwave carrier
frequencies and smaller cell size radius. The results show that for both
distance-dependent: path loss, shadowing and eective road height the
uplink information capacity of the cellular wireless system decreases
as carrier frequency f
c
increases and cell size radius R decreases.
For example at a carrier frequency f
c
= 15.75 GHz, basic path loss
exponent = 2 and cell size radius R = 100, 500 and 1000 m the
decrease in information capacity was 20, 5.29 and 2.68%.
1. INTRODUCTION
The common approach for utilizing the radio spectrum eciently is by
reducing the cell size radius of a cellular wireless network. Wireless
communication system have evolved from accommodating a large
number of voice and/or low speed data service. Now emerging and
future wireless system are required to accommodate high-speed data
Received 9 August 2012, Accepted 13 September 2012, Scheduled 24 September 2012
* Corresponding author: Kwashie Amartei Anang ([email protected]).
192 Anang et al.
service. Therefore it is important for wireless systems to achieve a high
degree of spectral eciency. Because the radio spectrum is limited
and regulated by international agreements [1, 2]. Takada et al. in [3]
reported that in order to achieve high data rates in wireless system,
broadband channels are required. Thus, carrier frequency needs to be
increased. References [46], also stated that for future mobile wireless
systems to meet the demand for high-speed data service they need to
be accommodated in frequency bands located above 2 GHz. However,
at higher carrier frequency cell size needs to be decreased because of
an increase in path loss [3].
For larger cell size radius (conventional mobile radio systems)
cellular systems, signal variability is characterized by Rayleigh
Statistics. Whilst for smaller cell size radius (higher frequency mobile
radio systems) cellular wireless system base station (BS) antenna
height are placed below the rooftops of surrounding buildings. Thus,
establishing a line-of-sight (LOS) radio link between the transmitter
and the receiver. This reduces the signal variability as compared to
the conventional mobile radio systems. Propagation characteristics
at higher carrier frequency are therefore dierent from conventional
cellular wireless system. With the rapid development of high-speed
data services in wireless communication systems, together with the
limited radio spectrum it is imperative that the eect of propagation
model and system parameters on the performance of a cellular wireless
network are studied in a comprehensive manner. Because propagation
loss is a fundamental phenomenon which aects the performance of
any wireless network.
The impact of propagation loss and system parameters on
information capacity performance of land mobile cellular wireless
systems has been studied and published in previous literature [617].
However, most of this studies have been performed in situations where
carrier frequencies were below 2 GHz. Co-channel interference were
considered to be from the rst tier and propagation characteristics
were based on the single-slope path loss propagation model.
Cox et al. in [7] performed an experiment to show the dependence
of wireless signal attenuation on the antenna height, however the
carrier frequency used for the experiment was 800 MHz. Green in [8],
carried out signal and path loss variability analysis for small cells and
concluded that path loss characteristics within smaller cell consist of
two distinct regions, however the analysis was for carrier frequency
below 2 GHz. Cruz-Perez and Lara-Rodrguez in [9] studied the
impact of breakpoint distance on the capacity of Manhattan like
microcellular CDMA system, without considering hexagonal cell layout
and interference were from the rst tier only. In [10], Min and Bertoni
Progress In Electromagnetics Research B, Vol. 44, 2012 193
studied the eect of path loss model on CDMA system design for
highway microcells, however they considered a highway scenario where
interference came from the rst tier.
Alouini and Goldsmith studied the eect of normalized reuse
distance, path loss, cell radius and shadowing on the capacity of
cellular wireless network using the area spectrum eciency [11].
However, they considered carrier frequencies f
c
= 900 MHz and
2 GHz, without considering carrier frequency greater than 2 GHz,
eective road height and the second tier co-channel interfering cells.
Har and Bertoni studied the eect of cell parameters on microcell
planning, considering the rst and second tier without considering
carrier frequency greater than 2 GHz [14]. Hernandez et al. studied
the eects of cell radius reduction and path loss model parameters
on the performance of microcellular networks without the second tier
co-channel interfering cells [13]. In [6], Hernandez et al. studied the
sensitivity of system performance to propagation parameters in LOS
microcellular environment for carrier frequencies greater than 2 GHz,
using outage probability for the performance analysis and considered
only urban and cigar-shaped microcellular environment.
Anang et al. in [16, 17] studied the impact of base station antenna
height and path loss exponent on the uplink information capacity of
a cellular wireless system operating at carrier frequencies greater than
2 GHz, when second tier co-channel interfering cells are considered to
be active. However, without considering the eect of shadowing and
eective road height.
Most of the previous studies were limited to cellular wireless
systems operating at the carrier frequencies f
c
< 2 GHz. In the case
where f
c
> 2 GHz, co-channel interference were assume to be due to
the rst tier only and not from any other tier. This assumption was
founded on large path loss exponent [18]. However, at higher carrier
frequencies cell size radius reduces and path loss exponent becomes
smaller, thus interference from the second tier becomes active [19].
Ge et al. in [20], reported that co-channel interference modeling and
capacity analysis are of importance in the next generation mobile
communication. However the interference model proposed was for
multi-cell multi-antenna cellular wireless communication system.
Many propagation models have also undergone corrections in
recent times, to suit other criterias such as rough terrain areas,
orchard environment, over water and wideband Digital Terrestrial
Television [2124]. Therefore, there is a need to understand the eect
of system and propagation loss parameters on the performance of
emerging and future cellular wireless system, which will be operating
at carrier frequencies greater than 2 GHz, where cell size radii are
194 Anang et al.
smaller and second tier co-channel interfering cells are active. The
main contribution of this paper is as follows.
We proposed an improved co-channel interference model, which
includes the second tier co-channel interfering cells.
We analyzed the impact of system parameters such as antenna
height and cell size on the uplink information capacity of a cellular
wireless system operating at carrier frequency greater than 2 GHz,
when both the rst and second tier co-channel interfering cells are
active.
We studied the impact of propagation loss parameters such as
shadowing component, path loss and eective road height on the
uplink information capacity of a cellular wireless system operating
at carrier frequency greater than 2 GHz, when both the rst and
second tier co-channel interfering cells are active.
The rest of the paper is organized as follows. Section 2 describes
the system model for propagation, co-channel interference and user
position, and outlines the basic assumptions used in modelling. In
Section 3, we focus on the average area spectral eciency of the land
mobile cellular system, which we use for our information capacity
performance measure. Section 4 presents some analysis and numerical
results. Finally, in Section 5, some conclusions are drawn from the
numerical results.
2. PROPAGATION AND SYSTEM MODELS
We considered a two-dimensional hexagonal microcellular wireless
network, where base stations (BSs) are uniformly distributed. Cells
form clusters (co-channel cell) around reference cells (BS
0
) as shown in
Fig. 1, and BSs are located at the centre of the cells. Each cell receives
signals from all users in the system which are attenuated according to
the power-law path loss and shadowing.
2.1. Microcellular Propagation Model
The radio environment of a cellular system is described by: (1)
path loss, (2) small-scale fading and (3) large-scale fading. In this
work, only path loss and shadowing are considered as we assume a
scenario where an ecient antenna diversity combining system is used
at the BS to eliminate the eects of multipath fading. Path loss
requires an accurate modeling, because it is one of the major sources
of systems performance degradation in cellular wireless system [25].
LOS propagation is assumed. Because users in the non LOS condition
(a) (b)
Figure 1. Hexagonal cellular site layout. Showing co-channel
interfering cells. (a) Conventional cellular site layout with 6 co-channel
interfering cells and reuse pattern (N = 7). (b) Cellular site layout
with 6 and 12 co-channel interfering cells, and reuse pattern (N = 7).
represents less than 1% of the LOS interference, hence non LOS can
be neglected [26]. Where there exist a LOS propagation between
transmitting and receiving antennas radio wave propagation modeling
is best done by the two-slope path loss model [2729]. Therefore, in
this paper we adopted the two-slope path loss model for our analysis
and simulations, which is given by:
S
r
=
K
r
(1 +r/g)
S
t
, (1)
where K, is the constant path loss factor, and it is the free space
path loss at a reference distance, d
0
= 1 m. is the basic path
loss exponent (ranges from 24) for r g and the additional path
loss exponent (ranges from 28) for r > g. The parameter r [m] is
the separation distance between transmitter and receiver, S
r
[W] the
received signal power, and S
t
[W] the transmitted signal power. The
parameter g [m] is the breakpoint of the path loss curve and is given
by g = 4h
b
h
m
/
c
, where h
b
[m] is BS antenna height, h
m
[m] the
mobile station (MS) antenna height, and
c
[m] the microwave carrier
frequency wavelength [30]. We use typical values in [4, 31, 32] for the
analysis in this paper. Table 1 shows that as f
c
increases g becomes
longer, likewise when h
b
increases.
Figure 2, shows the average received signal power given by (1)
196 Anang et al.
Table 1. Calculated breakpoint distance for dierent carrier
frequencies.
Carrier
frequencies (f
c
)

c
=
310
8
f
c
h
m
g for
h
b
= 10 m
g for
h
b
= 15 m
900 MHz 0.33 m 1.8 m 216 m 324 m
2 GHz 0.15 m 1.8 m 480 m 720 m
3.35 GHz 0.09 m 1.8 m 804 m 1206 m
8.45 GHz 0.035 m 1.8 m 2028 m 3042 m
15.75 GHz 0.019 m 1.8 m 3780 m 5670 m
1 Free Space Path Loss Model
2 Power falls off 1/(distance)
4
3 Two Ray Model
Figure 2. Two-slope path loss model; received signal strength vs.
distance, showing breakpoint distance, g for line of sight propagation.
(Transmit power, S
t
= 1 W, transmit and receiver antenna gains;
G
t
and G
r
= 0 dB, path loss exponent; = 2, extra path loss
exponent, = 4, base and mobile station antenna heights; h
b
= 15 m
and h
m
= 1.8 m, reference distance d
0
= 1 m and carrier frequency,
f
c
= 15.75 GHz carrier frequency.
plotted as a function of distance for MS and BS antenna height of 1.8
and 15 m, and carrier frequency of 15.75 GHz. The gure shows that for
the path loss, two regions may be distinguished, which are separated
by the breakpoint g. Before g the path loss is similar to free-space
path loss (path loss decreases 1/(distance)
2
). The path loss after g
decreases in inverse proportion to the fourth power of distance (path
loss decreases 1/(distance)
4
).
In the vicinity of high obstacles particularly in urban microcellular
systems the propagation is also aected by the shadowing of the LOS
path from the terrain, buildings and trees. Shadowing is modeled
as lognormal distribution [33]. The Probability Distribution Function
(PDF) of the slowly varying received signal power is thus given by the
standard lognormal expression as
p(P) =

2
P
P
exp
_
( ln(P)
P
)
2
2
2
P
_
, P 0 (2)
where = 10/ ln10 4.3429 is a scaling constant, and
P
[dB] and
P
[dB] are the mean and standard deviation of ln(P).
2.1.1. Modied Breakpoint Distance
It was concluded in [27], that at higher carrier frequencies greater than
2 GHz (3.35, 8.45 and 15.75 GHz), for LOS path loss characterization
in terms of breakpoint distance, the breakpoint distance was found to
be dierent from that predicted by the rst Fresnel zone model in [30].
The breakpoint distance g
h
, for path loss characterization at carrier
frequencies greater than 2 GHz is therefore given by [4], as
g
h
=
4(h
b
h) (h
m
h)
c
, (h < h
m
) (3)
where h is the eective road height, which is less than h
m
. The
dierence between g for carrier frequencies less than 2 GHz and g
h
for
carrier frequency greater than 2 GHz is the result of change in h, which
is caused by vehicles, pedestrians, and other objects on the road. The
eective road height h, therefore depends on the trac on the road,
and it is estimated by minimizing the dierence between the estimated
and theoretical breakpoint distances. For light vehicular trac h, is
between 0.230.74 [m], and for heavy vehicular trac between 1.29
1.64 [m] [32].
Table 2. Breakpoint distance for light vehicular trac.
Carrier
frequencies (f
c
)

c
=
310
8
f
c
h
m
h
g
h
for
h
b
= 10 m
g for
h
b
= 15 m
3.35 GHz 0.09 m 1.8 m 0.3 649.9000 m 984.9000 m
8.45 GHz 0.035 m 1.8 m 0.3 1639.3 m 2484.3 m
15.75 GHz 0.019 m 1.8 m 0.3 3055.5 m 46305 m
198 Anang et al.
Kitao and Ichitsubo in [5] stated that at higher carrier frequencies,
because ranges from 2 to 9 cm, which are smaller compared to the
size of vehicles. Therefore it can be assumed that the surface of
the road becomes uniformly higher and reection mainly caused by
vehicles can be handled by h. Experimental results in [4] show that
the breakpoint distance measured at night is more than that measured
during the day, because at night there is a lower amount of trac on
the road. Table 2 shows g
h
for dierent f
c
and h
b
for light vehicular
trac. Table 3 shows g
h
for dierent f
c
and h
b
for heavy vehicular
trac. Figure 3 reveals that as the eective road height increases the
Table 3. Breakpoint distance for heavy vehicular trac.
Carrier
frequencies
(f
c
)
c
=
310
8
f
c
h
m
h
g
h
for
h
b
= 10 m
g for
h
b
= 15 m
3.35 GHz 0.09 m 1.8 m 1.29 198.4138 m 312.3138 m
8.45 GHz 0.035 m 1.8 m 1.29 500.4766 m 787.7766 m
15.75 GHz 0.019 m 1.8 m 1.29 932.8410 m 1468.3 m
1 Carrier frequency, (f
c
= 15.75 GHz)
c
= 8.45 GHz)
c
= 3.35 GHz)
c
= 2 GHz)
c
= 900 MHz)
Figure 3. Breakpoint distance, g vrs. eective road height, h for
dierent carrier frequencies f
c
. BS and MS antenna height; h
b
= 15 m
and h
m
= 1.8 m.
breakpoint distance becomes shorter. It also shows that as carrier
frequency increase breakpoint distance becomes longer.
2.2. Improved Co-channel Interference Model
The rst and second tiers of co-channel cells are considered in
interference generation, unlike the existing model ([11]), where only
rst tier co-channel cells are considered. The desired MS is located
in the central cell and the interfering mobiles are in cells in the rst
and second tiers. Fig. 1(a) shows a conventional cellular system layout
with 6 co-channel interfering cells surrounding the desired BS (BS
0
).
Fig. 1(b) shows a cellular layout with the rst 6 co-channel interfering
cells and the second 12 co-channel interfering cells surrounding the
desired BS, BS
0
.
To simplify the analysis the following assumptions have been made
in the co-channel interference model. First the system is considered
to be interference-limited, with thermal noise power negligible relative
to the co-channel interference power [34]. Thus, the ratio of carrier to
noise CNR, reduces to the carrier-to-interference power ratio CIR. All
inter-channel interference are considered to be negligible [34]. All BSs
are assumed to transmit the same power.
Reference [11], provides a formula for calculating the uplink
desired user CIR for traditional cellular wireless communication
systems. Now it is reported in [35] that the uplink interference at
a BS is the non-coherent sum of interference signals from the users
served by the BS and the users served by other BSs. Likewise the
desired user CIR, , is dened as the ratio of the average received
signal power from a MS at a distance r [m] from the desired BS to the
sum of interfering received signal power. Thus, the improved desired
user CIR, , can be written as follows:
=
S
r
S
I
=
S
r
(r)
N
FI
fi=1
S
fi
(r
fi
) +
N
SI
si=1
S
si
(r
si
)
, (4)
where S
r
[W] is the average received signal power level and S
I
[W] the
sum of interfering received signal power. N
FI
and N
SI
are the total
number of co-channel interfering cells in regions I & II (refer to Fig. 2),
which are assumed to be interfering cells in the rst and second tier.
S
fi
[W] and S
si
[W] are the average received signal power level from
the ith interfering MSs at distances r
fi
and r
si
in the regions I & II
from the desired BS.
200 Anang et al.
2.3. Trac Distribution
The cell shape is approximated by a circle of radius R. All MSs
(desired and interfering) are assumed to be uniformly and independent
distributed in their cells. MSs are also assumed to be located in the
far eld region. The probability distribution function (PDF) of a MS
location relative to its BS in polar co-ordinate is given by:
p
r
,
(r, ) =
(r R
0
)
(R R
0
)
2
; R
0
r R, 0 2. (5)
R
0
corresponds to the closet distance that a MS can be from the BS
antenna, (to be in the far eld region) and is approximately 20 m for
microcellular systems.
3. AREA SPECTRAL EFFICIENCY
The average area spectral eciency (ASE) measured in [bits/s/Hz/km
2
]
is dened as the achievable sum rate (of all users in a cell uplink) per
unit bandwidth per unit area and is given by [11] as:
A
e
=
N
s
k=1
C
k
(D/2)
2
(6)
where is the total bandwidth allocated to each cell, D the reuse
distance, and N
s
the total number of active serviced channels per cell.
The achievable sum rate C
k
[bits/sec] is the Shannon capacity of the
kth user, which depends on , the received carrier to interference power
ratio CIR of that user, and
k
the bandwidth allocated to the user. The
Shannon capacity formula (C = log(1 +CIR)) assumes interference
has Gaussian characteristics. Because both the interference and signal
power of the kth user vary with mobiles locations and propagation
conditions, varies with time, therefore the average channel capacity
of the kth user is given by [11] as:
C
k
=
k
_
+
0
log
2
(1 +) p
() d, (7)
where p
(), is the probability distribution function (PDF) of the

average mean CIR () of the kth user.
The transmission rate is assumed to be continuously adapted
relative to the CIR in such a manner that the BER goes to zero
asymptotically. In (7) assuming that all users are assigned the same
bandwidth, C
k
= (C), becomes the same for all users, therefore
A
e
can be written as:
A
e
=
4N
s
C
D
2
=
4N
s
C
R
2
u
R
2
, (8)
where R
u
is dened as the normalized reuse distance and is given by the
ratio of reuse distance and cell radius (D/R). For a FDMA systems,
where all users are allocated the same bandwidth
k
=
0
= /N
s
.
Substituting /N
s
in (7) followed by substitution into (8) yields:
A
e
=
4
R
2
u
R
2
_
+
0
log
2
(1 +) p
() d. (9)
For TDMA systems, the total bandwidth is allocated to one user
per time slot, thus N = 1, and
k
= . Substituting this in (8), shows
that TDMA systems have the same ASE as FDMA systems.
4. SYSTEM AND PROPAGATION LOSS PARAMETERS
IMPACT ANALYSIS
In this section, we compute the eect of both system and propagation
loss parameters on the ASE of a smaller cell size radius cellular wireless
system operating at microwave carrier frequency greater than 2 GHz,
in the presence of the rst and second tier co-channel interfering cells.
The analysis applies to a TDMA system, and is based on fully loaded
non-sectorized cellular systems ignoring multipath fading. Although
there is an excessive demand to broadcast (downlink) high speed data
in most emerging communication services, we focus on the uplink
between a MS and an intended BS. However, our analysis can be
extended to the downlink of a cellular wireless system. The analysis
presented in this section considers each BS to have an omnidirectional
antenna; however, it is straightforward to extend this to the case of
sectorization antennas.
4.1. Eect of Path Loss
As users are randomly located in their respective cells, the desired user
CIR , becomes a random variable depending on the position of the
(N
TI
+ 1) desired and interfering MSs, where N
TI
is the sum of N
FI
and N
SI
. The problem is reduced from (N
TI
+ 1) dimensions to 1
dimension to simplify the analysis by computing the corresponding
area spectrum eciency for the average case interference conguration.
With no power control all the N
FI
and N
SI
, co-channel interferes are
at the centre of their respective cells (average case interference), at
202 Anang et al.
distance r
fi
= D[m] and r
si
= 2D[m] from the desired mobiles
BS, where D denote the distance between BS
0
and rst tier BS
and 2D the distance between BS
0
and second tier BS [35]. Note
that power control is essential for direct sequence CDMA systems.
However in this paper we consider the most representative cellular
technology, which is TDMA based. Fig. 1(b) shows the cellular system
conguration. Assuming the transmitted power of all users are the
same and substituting (1) into (4) yields:
(r, N
FI
, N
SI
)
=
S
r
(r)
N
FI
fi=1
S
fi
(r
fi
) +
N
si
si=1
S
si
(r
si
)
=
r
(1 +r/g)
N
FI
fi=1
r
fi
(1 +r
fi
/g)
+
N
SI
si=1
r
si
(1 +r
si
/g)
=
r
(1 +r/g)
N
FI
fi=1
RRu
(1 +RR
u
/g)
+
N
SI
si=1
(2RR
u
)
(1 + (2RR
u
)/g)
=
_
2
N
FI
+N
SI
N
FI
N
SI
_
_
R
u
R
r
_
_
g + 2R
u
R
g +r
_
, (10)
where R
u
is the normalized reuse distance and R is the cell size radius.
Since is a function of r, the desired user capacity is given by:
C(N
TI
, r) = B
o
log
2
(1 +(N
FI
, N
SI
, r)). (11)
for given by (10). Substituting (11) in (8) yields the ASE
conditioned on the desired mobile position r, for a fully-loaded system.
Integrating (11) over the desired users position PDF (5) yields the
average ASE for the average interference conguration as:
A
e
(N
TI
) =
4
R
2
u
R
2
_
R
R
0
log
2
(1 +(N
FI
, N
SI
, r) p
r
(r)dr. (12)
It is clear from (12), that the average ASE depends on the area
mean, CIR which is a function of random locations of the MS, which
makes it mathematically intractable to solve. A computer simulation
is therefore used to solve it.
4.2. Eect of Shadowing
In this subsection, we compute the eect of shadowing on uplink
information capacity of a cellular wireless system operating at higher
microwave carrier frequencies greater than 2 GHz and smaller cell size
radius when both the rst and second tier co-channel interfering cell
are active. We consider a fully loaded cellular system where both
desired and interfering users are aected by lognormal shadowing
superimposed on path loss. This case do apply to BS radio receivers
which are able to average out the fast multipath fading, in this case
the system only react to lognormal channel variation. All assumption
made early in this paper still holds unless otherwise stated.
4.3. Analyses
We assumed the desired users signal to be lognormally shadowing
according to (2), (Subsection 2.1) with mean power
P
and standard
deviation
P
. There are N
TI
mutually independent lognormally
shadowed co-channel interferes (independent identically distributed
(i.i.d.)), each with mean
i
and standard deviation
i
.
4.3.1. Probability Distribution Function of Desired User
Carrier-to-interference Ratio
The total interference power S
I
is the sum of N
TI
i.i.d. lognormally
distributed random variables (RV) in the rst and second tier (N
FI
and
N
SI
). Generally closed-form expressions for the PDF of lognormal
sum are not known, however it has been accepted that it can be
approximated by another lognormal distribution. In this work, we
use the Fenton-Wilkinson method [33], because of its simplicity. Using
this method, the logarithmic mean
P
I
and logarithmic variance
P
I
of S
I
can be approximated by matching the rst- and second-order
moments as follows:
P
I
=
I
+ ln(N
TI
) +

2
I
2

2
ln
_
(N
TI
1) +e
2
I
/
2
N
TI
_
(13)
where N
TI
is the sum of co-channel interfering cells in the rst and
second tier, which in this case is 18. Thus (13) becomes
P
I
=
I
+ ln(18) +

2
I
2

2
ln
_
17 +e
2
I
/
2
18
_
(14)
and
2
P
I
=
2
ln
_
17 +e
2
I
/
2
18
_
(15)
Since the ratio of two lognormal RVs is also a lognormal RV, it follows
that the PDF of the desired MS CIR can be approximated by a
204 Anang et al.
lognormal PDF
p
d
(
d
) =

d
exp
_
( ln(
d
)
d
)
2
2
2
d
_
,
d
0 (16)
whose logarithmic mean and variance are
d
=
P

P
I
(17)
d
=
2
P
+
2
P
I
(18)
Now inserting (16) in (7) yields the desired MSs average capacity as
follows:
C =
W
o
log
2
(e)
d
_
+
0
ln(1 +
d
)
d
exp
_
( ln(
d
)
d
)
2
2
2
d
_
d
d
(19)
Substituting (19) in (8) yields the ASE for the average case interference
conguration in a shadowing environment as
A
e
(r) =
4C(r)
W
o
R
2
u
R
2
, (20)
The mean in (17), is computed as follows
d
= ln
_
2
_
R
u
R
r
_
_
g
2
+ 3gR
u
R + 2R
2
u
R
2
g +r
_
1
2
(g + 2R
u
R)
+ (g +R
u
R)
_
ln(18) +

2
PI

2
I
2
(21)
and
2
d
is given by (18). As the expressions in (20) is conditioned on
the desired MS position, the site specic average ASE at a distance r
and angle from the central base station, BS
0
is given by:
A
e
=
4
R
2
u
R
2
_
R
R
0
log
e
(e)
_
+ exp
_
+

2
d
2
2
__
p
r
(r)dr (22)
Due to the random location of MSs in their cells, (22) is mathematical
intractable to solve explicitly.
4.4. Computer Simulations
As A
e
appears to be mathematically intractable to solve explicitly
analytically, a Monte Carlo simulation is used to estimate it. The basic
parameter used for the simulations are summarized in Table 4, and
Fig. 4, shows a simplied owchart for the simulation. The simulation
methodology is as follows.
Figure 4. ASE simulation owchart for distance-dependent path loss.
4.4.1. Path Loss
I. One MS is randomly located within each cell, according to uniform
distribution as described in Subsection 2.3 of Section 2.
206 Anang et al.
Table 4. Simulation parameters.
Parameters Values
Cell radius R 100 to 1000 m
Path loss exponent () 2, 2.5, 2.8 and 3
Additional path loss exponent () 4
Reference distance (d
0
) 1 m
Path loss constant (K) 20 log
10
(
c
/4d
0
) dB
MS transmission power (S
t
) 1 W
BS antenna height (h
b
) 10 and 15 m [31]
MS antenna height (h
m
) 1.8 m [31]
Mobile Distribution Uniform and Random
Interference First and Second Tiers
Carrier frequencies f
c
0.9, 2, 3.35, 8.45
and 15.75 GHz [4, 11]
Frequency reuse factor 4 [11]
II. The polar coordinates (x
fi
,
fi
and x
si
,
si
) of the N
FI
and N
SI
co-channel interferers are randomly picked according to (5).
III. We represent the location of the MSs and the BSs using
rectangular and polar coordinate system. Distances between MSs
and BSs are computed, and nally distances between desired BS
0
with respect to the ith interfering MSs (see Fig. 5) are computed.
The distance r
fi
for each co-channel interferer from the rst tier
to the desired BS is calculated as,
r
fi
=
_
D
2
+x
2
fi
2Dx
fi
cos(
fi
) (23)
IV. The distance r
si
for each co-channel interferer from the second
tier to the desired BS is calculated as:
r
si
=
_
(2D)
2
+x
2
si
4Dx
si
cos(
si
) (24)
V. The dual-slope path loss model (1), is used to calculate the area
mean power of the desired user and interfering mobiles in the rst
and second tier of co-channel cells (S
d
, S
fi
s and S
si
s), therefore
CIR is calculated as.
=
1
r
(g +r)
_
N
FI
r
fi
=1
1
r
fi
(g+r
fi
)
+
N
SI
r
si
=1
1
r
si
(g+ri2)
_
(25)
Figure 5. Geometry of desired MS and interfering MSs in two co-
channel cells.
VI. The ASE, A
e
is calculated as:
A
e
=
4
R
2
u
R
2
log
2
(1 +). (26)
Repeating the proceed above (from steps 16) 100 000 times. The value
of A
e
was estimated by taking the average of all the observations of
A
e
as given by (26).
4.4.2. Shadowing
For the case of shadowing the algorithm described in Section 4.4.1 is
change from step 5) to incorporate the eect of shadowing as follows:
VII. The N
TI
+ 1 area mean powers (
P
and
i
s
) at the considered
BS are calculated using the dual-slope path loss model (1).
VIII. The instantaneous received power from the desired and
interfering mobiles (S
d
, S
fi
s and S
si
s) are generated randomly
according to a lognormal distribution (2) with area means
P
and
s
, and standard deviations
P
and
I
.
We repeat the proceed for the shadowing case 1000 000 times to average
out the additional eect of shadowing to obtain the same degree of
accuracy for the average ASE.
4.5. Numerical Results
This section present Numerical results. The aim of the results
presented in this section is to make clear the inuence of system
and propagation loss parameters on the performance of emerging and
208 Anang et al.
future cellular wireless systems, rather than the mathematical analysis
itself. It is important to note that the main contributions of our work
are the ndings we obtained by analyzing the impact of system and
propagation loss parameters on the information capacity performance
of cellular wireless system operating at higher carrier frequency f
c
>
2 GHz and smaller cell size radius, when the rst and second tier co-
channel interfering cells are active. Our simulation compares result
for the conventional interference model presented in [11] (rst tier co-
(a) (b)
(c) (d)
1 Single tier co-channel interfering model (f
c
= 900 MHz)
2 Two tier co-channel interfering model (f
c
= 900 MHz)
c
= 2 GHz)
c
= 2 GHz)
c
= 3.35 GHz)
c
= 3.35 GHz)
c
= 8.45 GHz)
8 T wo tier co-channel interfering model (f
c
= 8.45 GHz)
c
= 15.75 GHz)
c
= 15.75 GHz)
Figure 6. Average uplink area spectral eciency versus cell radius
for dierent path loss exponents, , and carrier frequencies f
c
. (Fully-
loaded system with 6 and 12 co-channel interfering cells in rst and
second tier N
FI
= 6 and N
SI
= 12; extra path loss exponent, = 4;
MS and BS antenna heights: h
m
= 1.8 m and h
b
= 10 m). (a) Path
loss exponent, = 2. (b) Path loss exponent, = 2.5. (c) Path loss
exponent, = 2.8. (d) Path loss exponent, = 3.
channel interfering cell only) to our propose interference model (rst
and second tier co-channel interfering cell) presented in this work (12).
Figure 6 shows the plot of ASE versus cell radius R for dierent
path loss exponents and carrier frequencies f
c
. As expected the
ASE improves as the path loss exponent increases, because the
interfering signals are more attenuated. This is in agreement with
existing work [11]. The graphs also quantied the fact that decreasing
cell size increases the ASE. The curves in Fig. 6(a) show that for
f
c
= 0.9 and 2 GHz, the second tier co-channel interfering cells have
no eect on the ASE as the curves for both the conventional and
proposed model overlay. For carrier frequency f
c
= 3.35 GHz and cell
Table 5. Percentage decrease in area spectrum eciency between the
two interference model.
h
m
= 1.8 m, h
b
= 10 m and = 4
Basic Path
Loss Exponent ()
Carrier
Frequency
Cell Size
Radius (m)
Percentage
Decrease (%)
2
900 MHz 100 3.96
2 GHz 100 6.80
3.35 GHz 100 9.68
8.45 GHz 100 15.85
15.75 GHz 100 19.47
2.5
900 MHz 100 2.78
2 GHz 100 4.68
3.35 GHz 100 6.59
8.45 GHz 100 10.84
15.75 GHz 100 13.46
2.8
900 MHz 100 2.29
2 GHz 100 3.8
3.35 GHz 100 5.25
8.45 GHz 100 8.62
15.75 GHz 100 10.68
3
900 MHz 100 2.00
2 GHz 100 3.27
3.35 GHz 100 4.57
8.45 GHz 100 7.4
15.75 GHz 100 9.17
210 Anang et al.
size radius R < 200 m there was a decrease of 9.7% in ASE between the
conventional and proposed model, whiles for R > 200 m there was no
decrease. When f
c
= 8.45 GHz for R < 500 m the decrease in ASE was
between 3.1316.51%, whiles for R > 500 m there was no much decrease
in ASE between the two models. For carrier frequency f
c
= 15.75 GHz
and cell size radius 100 m R 1000 m, there was a decrease in ASE
between 2.6520%. It can therefore be concluded that, for TDMA-
based cellular systems operating at higher carrier frequencies as cell
size reduces second tier co-channel interference becomes severe.
This behavior of decrease in ASE as carrier frequency increases
and cell radius decreases can be explained intuitively as follows; at
higher carrier frequency the breakpoint distance from the transmitter
h
m
= 1.8 m, h
b
= 10 m and = 4
Basic Path
Loss Exponent ()
Carrier
Frequency
Cell Size
Radius (m)
Percentage
Decrease (%)
2
900 MHz 500 0.8
2 GHz 500 1.12
3.35 GHz 500 1.52
8.45 GHz 500 3.14
15.75 GHz 500 5.28
2.5
900 MHz 500 0.52
2 GHz 500 0.71
3.35 GHz 500 1.0
8.45 GHz 500 1.96
15.75 GHz 500 3.31
2.8
900 MHz 500 0.44
2 GHz 500 0.6
3.35 GHz 500 0.78
8.45 GHz 500 1.5
15.75 GHz 500 2.45
3
900 MHz 500 0.4
2 GHz 500 0.48
3.35 GHz 500 0.67
8.45 GHz 500 1.23
15.75 GHz 500 2.07
becomes longer. Consequently, decreasing cell size will cause more co-
channel interfering cells to be before the breakpoint where signals are
attenuated according to the inverse square distance, thus decreasing
the ASE. Figs. 6(b), 6(c) and 6(d), show that as increases the
decrease in ASE between our model and conventional model reduces as
expected. Because interfering signals before and after the breakpoint
distance are more attenuated, even at higher carrier frequency. This
conrms the statement by Zhou et al. in [18], that in conventional
cellular system performance analysis, because large basic path loss
exponent are assumed second tier co-channel cell are neglected.
However this may not be the case for emerging cellular wireless network
as becomes smaller.
h
m
= 1.8 m, h
b
= 10 m and = 2
Extra Path
Loss Exponent ()
Carrier
Frequency
Cell Size
Radius (m)
Percentage
Decrease (%)
2
900 MHz 100 10.25
2 GHz 100 13.51
3.35 GHz 100 16.10
8.45 GHz 100 20.48
15.75 GHz 100 22.60
3
900 MHz 100 6.36
2 GHz 100 9.65
3.35 GHz 100 12.49
8.45 GHz 100 18.00
15.75 GHz 100 21.03
5
900 MHz 100 2.56
2 GHz 100 4.85
3.35 GHz 100 7.49
8.45 GHz 100 13.93
15.75 GHz 100 18.15
8
900 MHz 100 0.79
2 GHz 100 1.88
3.35 GHz 100 3.50
8.45 GHz 100 9.35
15.75 GHz 100 14.35
212 Anang et al.
The results of the percentage decrease in ASE between the two
interference model; for dierent basic path loss exponents , carrier
frequencies f
c
and cell size radii R; 100 and 500 m are given in Tables 5
and 6. Comparing the results for the conventional interference model
presented in [11] to our propose interference model presented in this
work (12), we can conclude that at higher carrier frequencies, smaller
path loss exponent and smaller cell radius, second tier co-channel
(a)
(b)
(c)
(d)
c
= 900 MHz)
c
= 900 MHz)
c
= 2 GHz)
c
= 2 GHz)
c
= 3.35 GHz)
c
= 3.35 GHz)
c
= 8.45 GHz)
c
= 8.45 GHz)
c
= 15.75 GHz)
c
= 15.75 GHz)
for dierent extra path loss exponents, , and carrier frequencies, f
c
.
(Fully-loaded system with 6 and 12 co-channel interfering cells in the
rst and second tier; N
FI
= 6 and N
SI
= 12; path loss exponent,
= 2; MS and BS antenna heights: h
m
= 1.8 m and h
b
= 10 m).
(a) Extra path loss exponent, = 2. (b) Extra path loss exponent,
= 3. (c) Extra path loss exponent, = 5. (d) Extra path loss
exponent, = 8.
interference causes a decrease in the information capacity (ASE) of
a cellular wireless communication system.
Figure 7 indicates the eect of extra path loss exponent on
the performance of cellular wireless system operating at higher carrier
frequencies and smaller cell size radius. The curves show that as f
c
increases from 0.9 to 15.75 GHz, and cell radius decreases from 1000
100 m, the ASE increases exponential. The gure conrms that ASE
is an increasing function of both and R. Fig. 7(a), shows that for the
case when = 2 and f
c
= 0.9, 2 and 3.35 GHz. The curves show a
small decrease in ASE from R 300 m. However, for f
c
= 8.45, and
15.75 GHz, when R 500 m the decrease in ASE between our model
and existing model are bigger.
An interesting observation is that for R = 300 m the ASE at
8.45 GHz for the two tier model equals the 15.75 GHz single tier
model. This conrms the importance of the consideration of second
tier co-channel interference in cellular system information capacity
performance analysis at higher carrier frequencies and smaller cell size
radius. Fig. 7(b), shows that for the case when = 3 and f
c
= 0.9,
2 and 3.35 GHz. The curves for the ASE for the two models overlay
each other. However, for f
c
= 8.45, and 15.75 GHz, when R 500 m
the decrease in ASE between our proposed model and conventional
model are between 518%, and 21.747.36%. Figs. 7(c) and 7(d), reveal
c
= 900 MHz)
c
= 900 MHz)
c
= 2 GHz)
c
= 2 GHz)
c
= 3.35 GHz)
c
= 3.35 GHz)
c
= 8.45 GHz)
c
= 8.45 GHz)
c
= 15.75 GHz)
c
= 15.75 GHz)
Figure 8. Average uplink area spectral eciency versus cell radius for
dierent f
c
. (Fully-loaded system with 6 and 12 co-channel interfering
cells in the rst and second tier; N
FI
= 6 and N
SI
= 12; path loss
exponent, = 2 and extra path loss exponent = 2; BS and MS
antenna height, h
b
= 15 m and h
m
= 1.8 m).
214 Anang et al.
that as increases the second tier co-channel interference become less
severe. This contributes to decrease in ASE between the two models
being minimal. Tables 7 and 8 show the results of percentage decrease
in ASE between the two interference model; for dierent extra path
loss exponents , carrier frequencies f
c
and cell size radii R; 100 and
500 m.
Comparing Figs. 7(a) and 8, reveals the eect of BS antenna height
h
b
on the system information capacity performance at dierent carrier
frequencies. It can be observed that for h
b
= 15 m and = 2 for
f
c
= 0.9, 2 and 3.35 GHz the curves show more decrease in ASE from
R 300 m as compared to Fig. 7(a). This is because as h
b
increases
h
m
= 1.8 m, h
b
= 10 m and = 2
Extra Path
Loss Exponent ()
Carrier
Frequency
Cell Size
Radius (m)
Percentage
Decrease (%)
2
900 MHz 500 3.64
2 GHz 500 4.44
3.35 GHz 500 5.29
8.45 GHz 500 7.78
15.75 GHz 500 10.24
3
900 MHz 500 1.67
2 GHz 500 2.18
3.35 GHz 500 2.80
8.45 GHz 500 5.0
15.75 GHz 500 7.36
5
900 MHz 500 0.39
2 GHz 500 0.59
3.35 GHz 500 0.84
8.45 GHz 500 2.00
15.75 GHz 500 3.9
8
900 MHz 500 -
2 GHz 500 -
3.35 GHz 500 0.16
8.45 GHz 500 0.55
15.75 GHz 500 1.49
g increases causing the second tier interfering cell to be in the region
before g. For f
c
= 8.45, and 15.75 GHz, when R 500 m the decrease
in ASE between our model and the existing are much bigger than in
Fig. 7(a). Unlike the case when h
b
= 10, where at R = 300 m the ASE
at 8.45 GHz for the two tier model equals that of the 15.75 GHz single
tier model. For the case of h
b
= 15, it is not equal.
Figure 9 illustrates the eect of the eective road height on ASE
(a) (b)
(c) (d)
c
= 3.35 GHz)
c
= 3.35 GHz)
c
= 8.45 GHz)
c
= 8.45 GHz)
c
= 15.75 GHz)
c
= 15.75 GHz)
for light and heavy vehicular trac with eective road heights, h = 0.3
and 1.29 m and dierent carrier frequencies, f
c
. (Fully-loaded system
with 6 and 12 co-channel interfering cells in the rst and second tier;
N
FI
= 6 and N
SI
= 12; path loss exponent, = 2 and extra path loss
exponent = 2; MS antenna height, h
m
= 1.8 m). (a) Light vehicular
trac, BS antenna height, h
b
= 10 m. (b) Heavy vehicular trac, BS
antenna height, h
b
= 10 m. (c) Light vehicular trac, BS antenna
height, h
b
= 15 m. (d) Heavy vehicular trac, BS antenna height,
h
b
= 15 m.
216 Anang et al.
at carrier frequency greater than 2 GHz, when the rst and second tier
co-channel interfering cells are active. We found that previous result
in this paper still holds. This gure shows that the decrease in ASE
increases as carrier frequency increases and cell size radius decreases. It
is observed that the decrease in ASE is lesser for heavy vehicular trac
as compare to light vehicular trac. For the case of light vehicular
trac at R = 100 m, for f
c
= 3.35, 8.45 and 15.75 GHz the decrease in
ASE is 15.11, 19.55 and 21.94%, whilst for heavy vehicular trac it is
9.94, 13.69 and 16.61%. Then, we conclude that at microwave carrier
frequencies greater than 2 GHz and smaller cell size radius, the eect of
second tier co-channel interference is minimal during heavy vehicular
trac (day time) relative to light vehicular trac (night time). This
is because the breakpoint distance at day time is bigger than that at
night time. This is due to the increment of the eective road height
[refer to 3]. The gures show that increasing h
b
leads to second tier
co-channel interference becoming more severe. The gure also reveals
that even if h
b
is increased, the decrease in ASE is still lesser for day
time.
The eect of shadowing on the ASE at higher carrier frequency is
(a) (b)
c
= 900 MHz)
c
= 900 MHz)
c
= 2 GHz)
c
= 2 GHz)
c
= 3.35 GHz)
c
= 3.35 GHz)
c
= 8.45 GHz)
c
= 8.45 GHz)
c
= 15.75 GHz)
c
= 15.75 GHz)
in a shadow environment with dierent carrier frequencies, f
c
. (Fully-
loaded system with 6 and 12 co-channel interfering cells in the rst
and second tier; N
FI
= 6 and N
SI
= 12; path loss exponent, = 2
and extra path loss exponent = 2; MS antenna height, h
m
= 1.8 m).
(a) Shadowing standard deviation, = 8 dB. (b) Shadowing standard
deviation, = 12 dB.
shown in Fig. 10. The curves have the same relative shape as previous
curves in this paper. However, the decrease in ASE between the two
models is always bigger. This is because of the variability of the signal.
The gure shows that the decrease in information capacity between the
two model becomes bigger as the carrier frequency increases and cell
size radius decreases. Comparing Figs. 10(a) and 10(b) reveals that
as increases the second tier co-channel interference become severe.
Fig. 11, shows that even in a shadow environment the decrease in
ASE is still lesser for heavy vehicular trac compare to light vehicular
trac.
We see that in all cases the ASE curves for dierent interference
congurations have the same relative shape. Then from Figs. 9 and 11,
it is concluded that, for a TDMA-based cellular system operating at
higher carrier frequencies, for heavy vehicular trac the decrease in
ASE due to second tier co-channel interference is minimal relative to
the one obtained for light vehicular trac. This is in agreement with
results reported in [6]. However, in [6] the evaluation was based on the
outage probability, a value for radius and a cigar-shaped microcellular
(a) (b)
c
= 3.35 GHz)
c
= 3.35 GHz)
c
= 8.45 GHz)
c
= 8.45 GHz)
c
= 15.75 GHz)
c
= 15.75 GHz)
for light and heavy vehicular trac with eective road heights,
h = 0.3 and 1.29 m and dierent carrier frequencies, f
c
in a
shadowed environment. (Fully-loaded system with 6 and 12 co-channel
interfering cells in the rst and second tier; N
FI
= 6 and N
SI
= 12;
path loss exponent, = 2 and extra path loss exponent = 2; BS
and MS antenna height, h
b
= 10 m, and h
m
= 1.8 m and shadowing
standard deviation, = 8 dB). (a) Eective road height h = 0.3 m.
(b) Eective road height h = 1.29 m.
218 Anang et al.
scenario. The author found that the outage probability is improved
for heavy vehicular trac as compared to light vehicular trac.
In addition, we have found that second tier co-channel interfering
cells become severe as cell size radius reduces and carrier frequency
increases.
5. CONCLUSION
We studied and analyzed the impact of both system and propagation
loss parameters on the information capacity performance of a TDMA
based cellular wireless communication system operating at higher
carrier frequency and smaller cell size radius, when both rst and
second tier co-channel interference are active. Our studies and analysis
were based on the two slope propagation model characterized by
lognormal shadowing and path loss.
From the results, we can conclude that for higher carrier
frequencies and smaller cell size radius, second tier co-channel
interference becomes active. As a result, the information capacity
of the cellular wireless communication system is reduced. We also
conclude that as the basic path loss exponent decreases the second
tier co-channel interference becomes severe. Likewise when the
carrier frequency increases and cell size radius reduces. In the case
where shadowing is considered, we can conclude that the higher the
shadowing coecient the more severe is the second tier co-channel
interference. In future work, it may be interesting to study the eects
of sectorization, correlation coecient and multipath fading. Though
it may be expensive to carry out an experiment to compare simulated
results to experimental result, in future we may perform an experiment
for results comparison.
REFERENCES
1. Lee, W. C. Y., Spectrum eciency in cellular, IEEE Trans. Veh.
Technol., Vol. 38, 6975, May 1989.
2. Pahlavan, K. and A. H. Levesque, Wireless data communica-
tion, Proc. IEEE, Vol. 82, 13981430, Sep. 1994.
3. Takada, J., J. Fu, H. Zhu, and T. Kobayashi, Spatio-
temporal channel characterization in a suburban non line-of-sight
microcellular environment, IEEE J. Select. Areas Commun.,
Vol. 20, No. 3, 532538, Apr. 2002.
4. Masui, H., T. Kobayashi, and M. Akaike, Microwave path-loss
modeling in urban line-of-sight environments, IEEE J. Select.
Areas Commun., Vol. 20, No. 6, 11511155, Aug. 2002.
5. Kitao, K. and S. Ichitsubo, Path loss prediction formula for
microcell in 400 MHz to 8 GHz band, IET Electronics Letters,
Vol. 40, No. 11, 685687, May 2004.
6. Hernandez-Valdez, G., F. A. Cruz-Perez, and D. Lara-Rodrguez,
Sensitivity of the system performance to the propagation
parameters in LOS microcellular environments, IEEE Trans.
Veh. Technol., Vol. 57, No. 6, 34883508, Nov. 2008.
7. Cox, D. C., R. R. Murray, and A. W. Norris, Antenna height
dependence of 800 MHz attenuation measured in houses, IEEE
Trans. Veh. Technol., Vol. 34, No. 2, 108115, May 1985.
8. Green, E., Radio link design for microcellular systems, British
Telecom. Technology, Vol. 8, No. 1, 1990.
9. Cruz-Perez, F. A. and D. Lara-Rodriguez, Impact of the break-
point distance on the reverse link capacity of a CDMA man-
hattan like microcellular system, Proc. 5th IEEE Interna-
tional Symposium on Spread Spectrum Techniques and Applica-
tions (ISSSTA1998), Vol. 1, 262266, Sun City, South Africa,
Sep. 1998.
10. Min, S. and H. L. Bertoni, Eect of path loss on CDMA
system design for highway microcells, Proc. 48th IEEE Vehicular
Technology Conference (VTC1998), 10091013, Ottawa, Canada,
May 1998.
11. Alouini, M. and A. J. Goldsmith, Area spectral eciency
of cellular mobile radio systems, IEEE Trans. Veh. Technol.,
Vol. 48, No. 4, 10471065, Jul. 1999.
12. Ho, C., J. Copeland, C. Lea, and G. St uber, Impact of
the cell size on the cells Erlang capacity and call admission
control in the DS/CDMA cellular networks, Proc. 51st IEEE
Vehicular Technology Conference (VTC2000-Spring), Vol. 1, 385
389, May 2000.
13. Hernandez-Valdez, G., F. A. Cruz-Perez, and M. Lara, Impact
of the cell size and the propagation model parameters on
the performance of microcellular networks, Proc. 11th IEEE
International Symposium on Personal, Indoor and Mobile Radio
Communications (PIMRC2000), Vol. 1, 292296, London, UK,
2000.
14. Har, D. and H. L. Bertoni, Eect of anisotropic propagation
modeling on microcellular system design, IEEE Trans. Veh.
Technol., Vol. 49, No. 4, 13031313, Aug. 2000.
220 Anang et al.
15. Ahmed, B. T., M. C. Ramon, and L. de Haro Ariet, Capacity
and interference statistics of highways W-CDMA cigar-shaped
microcells (uplink analysis), IEEE Commun. Lett., Vol. 6, No. 5,
172174, Aug. 2002.
16. Anang, K. A., P. B. Rapajic, T. I. Eneh, and G. Oletu, Sensitivity
of information capacity of land mobile cellular system to the base
station antenna height at higher microwave frequencies, Proc. 3rd
International Conference on Computer Research and Development
(ICCRD2011), Vol. 3, 167172, Shanghai, China, May 2011.
17. Anang, K. A., P. B. Rapajic, T. I. Eneh, and B. Lawal,
Sensitivity of information capacity of land mobile cellular system
to propagation loss parameters at higher microwave frequencies,
Proc. 7th IEEE International Wireless Communications and
Mobile Computing Conference, 630635, Istanbul, Turkey,
Jul. 2011.
18. Zhou, S., M. Zhao, X. Xu, J. Wang, and Y. Yao, Distributed
wireless communication system: A new architecture for future
public wireless access, IEEE Commun. Mag., Vol. 41, 108113,
2003.
19. Anang, K. A., P. B. Rapajic, T. I. Eneh, and Y. Nijsure,
Minimum cell size for information capacity increase in cellular
wireless network, Proc. 73rd IEEE Vehicular Technology
Conference (VTC2011), 305311, Budapest, Hungary, May 2011.
20. Ge, X., K. Huang, C. X. Wang, X. Hong, and X. Yang, Capacity
analysis of a multi-cell multi-antenna cooperative cellular network
with co-channel interference, IEEE Trans. Wireless Commun.,
Vol. 10, No. 10, 32983309, Oct. 2011.
21. Nisirat, M. A., M. Ismail, L. Nissirat, and S. A. Alkhawaldeh,
A terrain roughness correction factor for Hata path loss model
at 900 MHz, Progress In Electromagnetics Research C, Vol. 22,
1122, 2011.
22. Phaebua, K., C. Phongcharoenpanich, M. Krairiksh, and T. Ler-
twiriyaprapa, Path-loss prediction of radio wave propagation in
an orchard by using modied UTD method, Progress In Electro-
magnetics Research, Vol. 128, 347363, 2012.
23. Sirkova, I., Propagation factor and path loss simulation
results for two rough surface reection coecients applied to
the microwave ducting propagation over the sea, Progress In
Electromagnetics Research M, Vol. 17, 151166, 2011.
24. Caluyo, F. S. and J. C. D. Cruz, Antenna characterization
and determination of path loss exponents for 677 MHz channel
using xed and portable digital terrestrial television, Progress In
Electromagnetics Research C, Vol. 29, 149161, 2012.
25. Neskovic, A., N. Neskovic, and D. Paunocc, Macrocell electric
eld strength prediciton model based upon articial neural
networks, IEEE J. Select. Areas Commun., Vol. 20, No. 6, 1170
1176, Aug. 2002.
26. Cruz-Perez, F. A., D. Lara-Rodriguez, and M. Lara, Full- and
half-square cell plans in urban CDMA microcellular networks,
IEEE Trans. Veh. Technol., Vol. 52, No. 3, 502511, May 2003.
27. Oda, Y., K. Tsuunekawa, and M. Hata, Advanced los path-
loss model in microcellular mobile communications, IEEE Trans.
Veh. Technol., Vol. 49, No. 6, 21212125, Nov. 2000.
28. Zhao, X., J. Kivinen, P. Vainikainen, and K. Skog, Propagation
characteristics for wideband outdoor mobile communications at
5.3 GHz, IEEE J. Select. Areas Commun., Vol. 20, No. 3, 507
514, 2002.
29. Sarkar, T. K., Z. Ji, K. Kim, A. Medouri, and M. Salazar-
Palma, A survey of various propagation models for mobile
communication, IEEE Antennas Propagat. Mag., Vol. 45, 5174,
Jun. 2003.
30. Green, E. and M. Hata, Microcellular propagation measurements
in an urban environment, Proc. 1st IEEE International Sym-
posium on Personal, Indoor and Mobile Radio Communications
(PIMRC1991), 324328, London, UK, Sep. 1991.
31. G. TR25.996, 3GPP SCM channel models, 3GPP TR25.996,
Vol. v6.1.0, Sep. 2003.
32. ITU, Propagation data and prediction methods for planning
of short-range outdoor radiocommunication systems and radio
local area networks in the frequency range 300 MHz to 100 GHz,
Recommendation ITU-R P.1411-1, iTU Radiocommunication
Assembly, 2012.
33. St uber, G. L., Principles of Mobile Communication, 2nd Edition,
98100, Kluwer Academic, Norwell, MA, 2001.
34. Lee, W. C. Y., Mobile Communication Design Fundamentals, 142,
John Wiley & Sons, New York, NY, 1993.
35. Singh, S., N. B. Mehta, A. F. Molisch, and A. Mukhopadhyay,
Moment-matched lognormal modeling of uplink interference with
power control and cell selection, IEEE Trans. Wireless Commun.,
Vol. 9, No. 3, 932938, Mar. 2010.

Progress in Electromagnetics Research B, Vol. 44, 191-221, 2012

Uploaded by

Copyright:

Available Formats

Progress in Electromagnetics Research B, Vol. 44, 191-221, 2012

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Progress in Electromagnetics Research B, Vol. 44, 191-221, 2012

Uploaded by

Copyright:

Available Formats

Progress In Electromagnetics Research B, Vol.

44, 191221, 2012

(), is the probability distribution function (PDF) of the

You might also like