See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/255670340

Planning of WiMAX and LTE Networks

Article

CITATIONS READS
8 413

2 authors, including:

Johnson Agbinya
Melbourne Institute of Technology
202 PUBLICATIONS   1,946 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Wireless Power Transfer View project

Book Series, Elsevier/CRC Press Publishers View project

All content following this page was uploaded by Johnson Agbinya on 13 May 2016.

The user has requested enhancement of the downloaded file.

 4
Planning of WiMAX and LTE Networks

Mehrnoush Masihpour and Johnson I Agbinya

University of Technology, Sydney Australia

PART III
WiMAX and LTE Link Budget

4.1 Propagation Regimes

Network planning provides information on the feasibility of network role out.
Specifically some of the information is on the estimate of the optimum number
of base stations, the location of the base stations and antennas, determining the
type of the antenna and also the received power at receiver and the environment
characteristics of the propagation environment. These optimum values are
dependent on the services and the number of users.
In wireless systems, the transmitted signal is propagated in an open envi-
ronment and losses some of its power as a result of phenomenon such as
scattering, diffraction, reflection and fading when received at the receiver;
therefore it is important to calculate the received signal power to determine
factors such as the radius of a cell or the type of the antenna [1]. The degra-
dation of the transmitted signal through interaction with the environment is
known as path loss, which means the difference between the transmitted and
received signal power. In general, path loss is calculated by [1]:
Path Loss = PT + GT + GR − PR − LT − LR [dB] (4.1)

105
106 Planning of WiMAX and LTE Networks

PT GT − LP G R − AM
PR

Fig. 4.1 Link budget.

Where, PT shows the power at transmitter and PR is power at the receiver, GT

and GR is the transmitter and receiver antenna gain respectively, LT and LR
express the feeder losses [1]. This equation describes the link budget. A link
budget describes the extent to which the transmitted signal weakens in the link
before it is received at the receiver. The link budget therefore accounts for all
the gains and losses in the path the signal takes to the receiver.
As shown in Figure 4.1, a link is created by three related communication
entities:

a) the transmitter
b) the receiver
c) the channel (medium) between them. The medium introduces
losses causing a resuction in the received power.

The link budget (Figure 4.1) equation therefore can be written either as in
equation (4.2) to show the path loss or in terms of the receiver power to be:
PR = PT + GT − LP + GR − AM (4.2)
This equation assumes that all the signal gains and losses are expressed in
decibels. The units for these are as follows:
AM (dB), (LT + LR in equation (4.2)) represents all the attenuation losses such
as feeder loss, link margin, diffraction losses, losses due to mobility (Doppler),
and the effects of rain, trees and obstacles in the signal path.
GT (dBi) is the transmitter antenna gain
GR (dBi) is the receiver antenna gain
PR (dBm) is the received power at the receiver
PT (dBm) is the transmitted power
LP (dB) is the path loss in the physical medium between the transmitter and
receiver.
 4.1 Propagation Regimes 107

Radio frequency sources are often modelled as isotropic sources. The

source radiates microwave energy uniformly in all directions, the so-called
isotropic source (into a spherical volume). Isotropic sources do not exist in
practice. A more practical source is the half-wave dipole that is used to model
an effective radiated power (ERP). Effective isotropic radiated power (EIRP))
is measured in terms of a half-wave dipole model and: EIRP = ERP + 2.15 dB
Interference margin is usually around 1 dB. This is used to account for
interference during the busy hour and depends on traffic load, frequency reuse
plan and other factors.
Penetration of microwave into buildings varies and can be very severe.
Building penetration is usually around 5 to 20 dB and accounts for penetration
into different types of building materials, for indoor coverage. Penetration is
a function of the type of building and the desired signal quality at the centre
of the building. In-building losses can be quite dramatic in areas such as lifts
and underground shelters.
Vehicle penetration is less severe compared to building penetration. The
estimate for this is around 6 dB and accounts for the attenuation of signal by
the frame of a car or truck. Other margins normally considered include:
Body loss (human body) of around 3 dB: This accounts for the absorption
of the signal by a mobile user’s head. It is sometimes called head loss. This
varies depending on hair structure and the proximity of the mobile terminal to
the head. Can you predict whether a female head will result to more body loss
than male? What about African and Caucasian head loss (these racial groups
carry different hair textures).
Fade margin of between 4 to 10 dB accounts for multipath fading as
discussed in Chapter 4. Fading dips occur for slow moving mobiles. Fast
moving mobiles tend to overcome this because they move faster out of dips
before the fading affects the signal. Fade margin varies with the environmental
conditions. In the next sections we describe the sequence of signal reducing
events in the
i) transmitter
ii) channel (medium) and
iii) receiver
The following discussions explain what happens at the transmitter, receiver
and channel respectively.
108 Planning of WiMAX and LTE Networks

GT
Antenna Gain

PT P = PT .GT

Base Station

Fig. 4.2 Base station antenna power gain.

Fig. 4.3 Sources of power losses in transmitter receiver.

4.2 Transmitter
A WiMAX base station normally employs some form of MIMO system and
the early implementations used 2 × 2 MIMO antennas. The base station also
used adaptive antenna systems. The transmitter antenna boosts the data signal
power before launching it into the channel.
The antenna creates an effective isotropically radiated power (EIRP) and
outputs (radiates) it into the channel (medium), where EIRP = ERP + 2.15 dBi.
Notice that the gain of the antenna is in dBi, while ERP is in dB. This con-
version is essential. Due to the connectors and cables (Figure 4.3) used in the
transmitter circuit, power losses are made as in Figure 4.4.
The antenna EIRP is reduced as follows

EIRP(dB) = PT − Lconnector,cable + GT (4.3)

 4.3 Signal Fading 109

Transmitter channel Receiver

GT
Antenna Gain Path Loss

PT P = PT .GT
Shadowoing

PR
RSS

Implementation margin

Base Station 
−102dBm Noise Figure
 KTB


Fig. 4.4 Received signal power level after passing through the channel.

Table 4.1 parmeters of base station [2].

BS with 2 × 2 MIMO and
Standard BS BS with 2 × 2 MIMO 2 element AAS
DLTx power 35 dBm 35 dBm 35 dBm
DLTx antenna gain 16 dBi 16 dBi 16 dBi
Other DLTx gain 0 dB 9 dB 15 dB
UL Rx antenna gain 16 dBi 16 dBi 16 dBi
Other UL Rx gain 0 dB 3 dB 6 dB
UL Rx noise figure 5 dB 5 dB 5 dB

For a 2 × 2 MIMO base station case, the following power losses were
recorded in [2].
The parameters for a set of customer premises equipment for both fixed
and mobile cases are given in Table 4.2.

4.3 Signal Fading

The WiMAX channel like all other communication channels introduces sig-
nal degradation (fading) and reduces the output signal launched to it by the
transmitter (Figure 4.4).
110 Planning of WiMAX and LTE Networks

Table 4.2 Parameters for customer premises equipment.

Portable CPE Mobile CPE
UL Tx power 27 dBm 27 dBm
UL Tx antenna gain 6 dBi 2 dBi
Other ULTx gain 0 dB 0 dB
DL Rx antenna gain 6 dBi 2 dBi
Other DLRx gain 0 dB 0 dB
DL Rx noise figure 6 dB 6 dB

The received signal power irrespective of the path loss and shadowing
must be greater than the WiMAX receiver sensitivity (RSS ). The signal to
noise ratio at the transmitter for a 2 × 2 MIMO base station is


FS Nused
SNR = PR + 102 − 10 log10 (4.4)
NFFT
This expression assumes that the implementation margin is 7 dB and noise
figure is 5 dB. The effects of the channel and all other degradation sources
must result to the received signal power level to be at least equal to or greater
than the receiver sensitivity PR ≥ RSS .
Signal degradation has always been a disturbing factor in telecommunica-
tions, and much more so in cellular communications. There are three princi-
pal sources of degradation in a cellular environment: noise (inter-modulation
noise, additive white Gaussian noise (AWGN)), multiple access interference
(MAI) and fading. Noise is mainly contributed by the environment and the
equipment in use. MAI is a result of sharing bandwidth and communication
channel. Multiple access interference refers to inter-cell interference, intra-
cell interference, co-channel interference and adjacent channel interference.
Inter cell interference exists between two or more cells because of shared
frequencies and frequency reuse. Intra-cell interference exists within a cell.
Co-channel interference occurs between two users using the same channel.
We have dealt with MAI in Chapter 3. The present chapter is dedicated to the
third type of impairment in cellular systems — fading.
Fading is of two forms, large-scale and small-scale fading. Large-scale
fading is mostly mean signal attenuation as a function of distance and signal
variation around its mean value. Small-scale fading is of two forms, time
spreading (which is composed of flat fading and frequency selective fading)
and time variance of channel (which is composed of fast fading and slow
fading). Figure 4.5 categorises these degradation sources.
 4.3 Signal Fading 111

Fig. 4.5 Impairment in cellular environment.

Ideally, we would like communication signals to propagate (travel) without

obstructions or disturbances. This can only happen in free space. Unfortu-
nately, free-space is not a practical proposition in telecommunication. Scat-
tering, reflection and diffraction of waves affect propagation of microwaves in
many ways. Multipath propagation, Doppler spread and coherence time, delay
spread and coherence bandwidth all conspire against the transmitted signal.
These effects reduce the system signal to noise ratio, and lead to incoherent
calls, dropped calls, and noisy channels.
Reflection occurs when the path of a wave is obstructed by a smooth
surface. As the wave impinges on the surface, its direction of travel is changed.
Usually the dimensions of the surface where reflection takes place are large
relative to the wavelength of the wave.
When an object with large dimensions relative to the wavelength of a signal
blocks the path of a wave, diffraction takes place. The object’s sharp edges
lead to diffraction or shadowing.
Scattering however occurs when a wave impinges on an object with dimen-
sions comparable to the wavelength of the signal. The object causes the wave
112 Planning of WiMAX and LTE Networks

to spread out or scatter to different directions. In urban applications street

lights, signs and foliage are the worst culprit. Scattering is a loss of useful
signal strength because the receiver is unable to collect all its energy as it is
tuned to a narrow view, and collects signals only in the narrow view.
What causes microwave signals to fade? The main causes of signal fading
are discussed in this section. Microwave signals arrive at a receiver from many
paths (multipath). The line of sight (LOS) signal is usually the preferred one
in most applications as it travels the shortest path and thus arrives at a receiver
with the strongest amplitude.

4.3.1 Multipath Propagation

Multipath propagation causes large variations in signal strength. The major
effects are threefold:

• Time variations due to multipath delays;

• Random frequency modulation due to Doppler shifts from different
multipath signals;
• Random changes in signal strength over short time periods;

In practice, multipath delays lead to time dispersion or ’fading effects’ which

is small-scale in nature.
A mobile communication multipath channel can be modelled as a linear
time-varying filter with impulse response h(t,τ ), where τ is the multipath delay
in the channel for a fixed time t. In practice a low-pass model of the channel is
easier to use than the actual complex model. The model allows a description of
all other signal components relative to the signal which arrives first with delay
τo = 0. All the delayed signals arriving latter than this are discretised in terms
of delays in N equally spaced time intervals of width τ . All the multipath
wave components in bin i are represented in terms of one component with
delay τι = iτ u
Multipath fading is measured by using channel sounding through direct
pulse measurements, spread spectrum sliding correlator or swept-frequency
channel analyser. These techniques provide time dispersion parameters (mean
excess delay, maximum excess delay at some given signal to noise ratio
and rms delay spread), coherence bandwidth and Doppler spread or spectral
broadening.
 4.3 Signal Fading 113

4.3.2 Doppler Effect

A moving object experiences Doppler frequency shifts. Mobile phone signals
in fast moving cars, air plane and ships also experience Doppler shifts. The
measured frequency increases as the mobile moves towards a base station. As
it moves away from the base station, the frequency decreases. Doppler effects
therefore leads to a variation of the signal bandwidth and is governed by the
expression:

v
fd = cos θ (4.5)
λ

where θ is the angle made by the signal path to the base station and the ground
plane as shown in Figure 4.6. In this expression, velocity v is measured in
meters/second, wavelength λ in meters and frequency fd in Hertz. In the time
domain, Doppler frequency shift leads to coherence time. Coherence time is
the time duration over which two signals have strong potential for amplitude
correlation. Coherence time can be approximated by the expression:

9
Tc = (4.6)
16πfm2

where fm is the maximum Doppler shift, which occurs when θ = 0 degrees. To

avoid distortion due to motion in the channel, the symbol rate must be greater
than the inverse of coherence time (1/Tc ).

θ d θ
A B
v

Fig. 4.6 Doppler effect in cellular communications.

114 Planning of WiMAX and LTE Networks

4.3.3 Delay Spread and Coherence Bandwidth

Multipath delay causes the signal to appear noise-like in amplitude. We can
compute its statistical averages and parameters. The standard deviation of the
distribution of multipath signal amplitudes is called delay spread, στ . Delay
spread varies with the terrain with typical values for rural, urban and suburban
areas: σt ≈ 0.2µs (rural); σt ≈ 0.5µs (suburban); σt ≈ 3.0µs (urban).
Since the signal bandwidth varies due to delays, what then is the best mea-
sure of the bandwidth in practice? To answer this question, we use coherence
bandwidth. It is defined to be the statistical measure of the range of frequen-
cies over which the channel is considered constant or flat. It is the bandwidth
over which two frequencies have a strong potential for amplitude correlation.
Coherence bandwidth estimated for both strong and weak correlation are:
Bc ≈ 0.02
σt
for correlation greater than 0.9 and Bc ≈ 0.2 σt
for correlation
greater than 0.5. Table 4.3 shows typical rms delay spreads for various types
of terrain.

4.3.4 Categories of Fading

There are two major categories of fading, small-scale and large-scale fading.
Large-scale fading is dependent on the distance between the transmitter and
receiver. It is generally called path loss or ‘large-scale path loss’, ‘log-normal
fading’ or ‘shadowing’. Small-scale fading is caused by the superposition of
multipath signals, the speed of the receiver or transmitter and the bandwidth
of the transmitted signal. Therefore, small-scale fading is a result of construc-
tive and destructive interference between several versions of the same signal
causing attenuation of the average signal power. This type of fading is over a

Table 4.3 Delay spread.

Delay spread figures at 900 MHz Delay in microseconds
Urban 1.3
Urban (worst-case) 10–25
Suburban (typical) 0.2–0.31
Suburban (extreme) 1.96–2.11
Indoor (maximum) 0.27
Delay Spread at 1900 MHz
Buildings (average) 0.07–0.094
Buildings (worst–case) 1.47
 4.3 Signal Fading 115

fraction of a wavelength and of the order of 20 to 30 dB. Small-scale fading

is also known by other names such as ‘fading’, ‘multipath’ and ‘Rayleigh’
fading. Rayleigh fading is a statistical variation of the received envelope of a
flat fading signal. Multipath fading manifests as time spreading of the signal
and a time variant behaviour. A time variant behaviour of the channel may
be due to motion of the mobile and or changing environment (movement of
foliage, reflectors and scatters).
If the impulse response of the mobile radio channel is h(τ , t) and time
invariant and if it has zero mean, then the envelope of the impulse response
has a Rayleigh distribution given by the expression:


r r2
p(r) = 2 exp − 2 (4.7)
σ 2σ
where σ 2 is the total power in the multipath signal. If however the impulse
response has a non zero mean, then there is a component of the direct path
(line of sight or specular component) signal in the channel and the magnitude
of the impulse response has a Ricean distribution. (Rice fading is therefore the
combination of Rayleigh fading with a significant non-fading (line of sight)
component). Ricean distribution is given by the expression

2   
r r + s2 rs
p(r) = 2 exp − 2
I0 2 (4.8)
σ 2σ σ
The power of the line of sight signal is s2 and I0 is a Bessel function of the
first kind.
The distance between either the dips or troughs in Rayleigh fading is of the
order of half a wavelength. Small-scale fading occurs as either of four types:

• frequency selective fading in which the bandwidth of the signal

is greater than the coherence bandwidth and the delay spread is
greater than the symbol rate; Signals at some frequency compo-
nents experience more fading than others
• flat fading when the bandwidth of the signal is less than the coher-
ence bandwidth and the delay spread is less than the symbol rate
• fast fading when the Doppler spread is high and the coherence time
is less than the symbol period and
• slow fading with a low Doppler spread and coherence time is greater
than the symbol period
116 Planning of WiMAX and LTE Networks

Table 4.4 Fading effects.

Type of fading Frequency Effects Time Effects
Effects of Multipath Delay Spread
Frequency Selective fading BW of signal > coherence BW Delay spread > symbol period
Flat fading BW of signal < coherence BW Delay spread < symbol period
Effects of Doppler Spread
Slow fading Low Doppler spread Coherence time > symbol period
Fast fading High Doppler spread Coherence time < symbol period

The first two fading are caused by multipath delay spread and the last two by
Doppler spread. Table 4.4 is a summary of the conditions that exist with each
type of fading.
Small-scale fading may be corrected by using adaptive equalisers or
through the use of modulation techniques such as spread spectrum and error
correction.

4.4 Propagation Regimes and Models

Based on different environment different path losses are experienced, con-
sequently different radio propagation modelling is employed to estimate the
electric field strength [1]. Channel models can be seen from different angles;
in terms of existence of a direct line of sight between the transmitter and
receiver, channel modelling methods can be categorized as free space prop-
agation model and modelling for the land propagation when the LOS does
not exist [3]. However, according to the required statistic for channel mod-
elling, two main groups of Deterministic and Empirical modelling can be
considered [1]. The deterministic model is used when the channel modelling
requires the detailed information about the geometric environment such as the
location and the surrounding area. In contrast, empirical models are based
on the measured mean path loss for different environments, this method
is less complex and cost but less accurate compared to the deterministic
modelling.
One of the empirical channels modelling usually is used by WiMAX
technology such as Stanford University Interim (USI) model, Cost-231 Hata
model, Macro Model and Ericsson model 9999 [1]. In this section the USI and
free space channel modelling are introduced in some detail.
 4.4 Propagation Regimes and Models 117

TX
RX

ht
d hr

Fig. 4.7 Two-ray propagation.

4.4.1 Terrestrial Microwave Propagation

Terrestrial propagation (propagation in natural settings) is much different from
free space propagation. Propagation of cellular signals in urban environment
is difficult to predict because of man made structures (tall buildings, signs
and other obstacles). On earth, a signal propagates through air and reflects,
diffracts and scatters as discussed earlier. Theoretically, propagation can be
predicted with the help of a 2-ray model and Fresnel Zones. A two-ray model
is shown in Figure 4.7. The signal reaches the receiver through two paths, the
direct path and a reflected ray. This model assumes a flat earth.
The path difference between the direct and reflected rays is given by the
expression:

   2ht hr
= (hr + ht )2 + d 2 − (ht − hr )2 + d 2 ≈ (4.9)
d
In practice, destructive earth reflections are avoided by imposing the criteria:

 > (2n − 1)λ/2, n = 1, 2, 3, . . . (4.10)

Electromagnetic wave fronts are often divided into zones of concentric cir-
cles (separated by half a wavelength) called Fresnel zones. The zones define
propagation break points. In the first Fresnel zone (no reflection) occurs and
n = 1 so that hr ht > λd/4. The first breakpoint, do is called the first Fresnel
zone and occurs at a distance do = 4hr ht /λ. Until this point, the propagation is
assumed to be free space. These two propagation models are useful for system
design. The two models are used to predict microcell and indoor coverage. In
many applications the distance between the transmitter and receiver is smaller
than the first breakpoint, and for such cases, the Fresnel point does not help
with the design.
118 Planning of WiMAX and LTE Networks

Table 4.5 Path loss exponent.

Environment Path loss exponent, n
Free space 2
Ideal specular reflection 4
Urban cells 2.7–3.5
Urban cells (shadowed) 3–5
In building (line of sight) 1.6–1.8
In building (obstructed path) 4–6
In factory (obstructed path) 2–3

Over ideal ground or the so-called specular ground, the received power is
given by the modified free-space power:


4π ht hr 2
Pr = PFS [(1 − exp(j 2π /λ))]2 ≈ PFS (4.11)
λd
h2r h2t
Pr = Pt Gt Gr (4.12)
d4
As a result of the distance dependence in this expression, every time we
double the distance, we lose 12 dB of signal energy. This shows that frequency
reuse should be done at shorter distances. The path loss exponent varies from
terrain to terrain as shown in Table 4.5.

4.4.2 Free Space Propagation

The medium separating the receiver from the transmitter plays an important
role in the propagation of the signal. Free space propagation is usually mod-
elled with the Friis formula [3]:
Gt Gr λ2
Pr = Pt (4.13)
(4π d)2
Where Pr , Pt , Gr and Gt are the receiver power, transmitter power, receiver
antenna and transmitter antenna gains respectively. It has been assumed in
this expression that radiation is into a spherical space of radius d surrounding
the antenna. Thus radiation is into an area equal to the surface area of a
sphere (4π d 2 ). The medium between the transmitter and receiver is often
a dielectric, air, a piece of wire, fibre or some liquid including water. The
medium varies from application to application and from terrain to terrain. The
medium between the transmitter and receiver introduces a propagation loss as
 4.5 Link Budget 119

Pt Pr

Lp

Gt Gr

Fig. 4.8 Free space propagation.

shown in Figure 4.8. The loss is modelled with the expression:


4π d 2
Lp = 10 × Log (4.14)
λ
The propagation-exponent is a function of the terrain between the transmitter
and receiver and has values in the range 2 to 5 in urban areas. A value 2 is
used for free space. The propagation loss is often expressed in decibels and is:

Lp = 32.44 + 20Log(f ) + 20Log (d) (4.15)

The frequency f is measured in MHz and the distance d in kilometres. As

the equation (4.14) implies, using higher carrier frequencies, the free space
loss will be higher. For instance, if two receivers have the same sensitivity,
using 450 MHz carrier frequency the cell radius will be 7.78 times longer than
when 3.5 GHz carrier frequency is used [4].

4.5 Link Budget

Propagation losses are useful for computing link budgets. Link budgets provide
the maximum allowable path losses per link and an indication of the link that
will likely be a limiting factor to the system. The maximum allowable path
loss also sets a limit on the maximum cell size. We will use path losses to
compute link budgets in the this chapter.
The plane earth model is more appropriate for cellular channels. The model
ignores the curvature of the earth’s surface and considers a two-path model
120 Planning of WiMAX and LTE Networks

of direct line of sight and a ground reflected paths. In this model the heights
of the transmitting and receiving antennae feature prominently in the propa-
gation loss expression. Provided the heights of the antennas are less than the
separation between the transmitter and receiver (hb and hr ), the propagation
loss expression can be shown to be

Lp = −10 log10 Gt − 10 log10 Gr − 20xLog hb

− 20xLog hr + 40xLog (d) dB (4.16)

4.5.1 Practical Models

Propagation out door is difficult to predict and as such, empirical models,
without real analytical basis are applied. Most of the models used are accurate
to within 10 to 14 decibels in urban and suburban areas. They tend to be
less accurate in rural areas because most of the data used may have been
collected in the urban and suburban areas. One of the other popular models
is the Okamura model. In practice there are huge variations in the types of
terrain and environment to cover. The heights of antenna, clutter, tree density,
beamwidth, wind speed, season (time of the year) and multipath, vary widely
and affect mobile phone waves. Hence complex models are required for such
situations. They are used to predict propagation loss. Beyond the Okamura
model, there are the Hata and Walfisch-Ikegami models and others by Egli,
Lee, Carey, Longley-Rice, Ibrahim-Parsons and many more. Although most
of them are beyond the scope of this note, we will examine a few briefly.
In a study of channel models for fixed wireless application by the Institu-
tion of Electrical and Electronic Engineers (IEEE) [5], a set of "propagation
models applicable to the multi-cell architecture" was presented. They assumed
a cell radius of less than 10 km, variety of terrain and tree density types, direc-
tional antenna (2–10 m) installed under-the-eaves/window or rooftop, 15–40
m BTS antennas and high cell coverage requirements of between 80–90%.
This section is derived from analysis.

4.5.1.1 Suburban Path Loss (Hata-Okamura Model)

The Hata-Okamura model [6] is the most widely used model for this situation.
The model is valid in the 150–1500 MHz frequency range (NMT and GSM),
with receiver distances greater than a kilometre from the base station and base
 4.5 Link Budget 121

Table 4.6 Path loss (Terrain) correction variables.

Model Parameter Terrain Type A Terrain Type B Terrain Type C
a 4.6 4 4.6
b 0.0075 0.0065 0.005
c 12.6 17.1 20

station antenna heights greater than 30m. As such, the model is applicable
to mobile phone applications below 1500 MHz. The modified Hata-Okamura
model extends this range to around 2 GHz [7]. This is the so-called COST 231
model. Although the model targeted 2G systems in the 900 and 1800 MHz
range, it has application to systems around 2 GHz (eg., DCS1800) provided
lower base station antenna heights, hilly or moderate-to-heavy wooded terrains
are not involved. Corrections for these limitations were applied to cover most
terrain conditions applicable to the US. The body of study for savannah, and
dense forest regions in Africa and other regions need to be understood. Path
loss regimes are divided into three broad categories A, B and C:

• Category A (maximum path loss): hilly terrain with moderate-to-

heavy tree densities
• Category B (intermediate path loss): terrain conditions between
category A and C
• Category C (minimum path loss): mostly flat terrain with light tree
densities

The median path loss at 1.9 GHz for a distance do from a base station is
given by:

Lp = A + 10 ∗ n ∗ Log10 (d/do ) + s; d > do (4.17)

where, A = 20xLog10 (4π do /λ) and λ is the wavelength in metres, n is the

path loss exponent and

n = (a − b hb + c/hb ) (4.18)

The height of the base station hb is between 10 m and 80 m, and do = 100 m, a,

b, and c are constants that depend on the terrain category which is reproduced
from [5] below.
The shadowing effect is represented by s and follows a log-normal dis-
tribution with typical standard deviation between 8.2 and 10.6 dB. This too
122 Planning of WiMAX and LTE Networks

depends on the terrain and tree density type. Correction terms are used to
account for antenna height and frequency region. For the model to apply to
frequencies outside the range of specification (2 GHz), and for receive antenna
heights between 2 m and 10 m, correction terms are specified. The coarse form
of the path loss model (in dB) has three correction terms as:

Lpc = Lp + Lf + Lh (4.19)

and Lf (in dB) is the frequency correction term given by the expression.
The frequency correction term is given by the expression

Lf = 6 Log(f/2000) (4.20)

The frequency f is MHz, and is positive for frequencies higher than 2 GHz.
The correction term for antenna height is:

Lh = −10.8 Log(h/2); for categories A and B and

Lh = −20 Log(h/2); for category C

The height of the receive antenna is in the range 2 m < h < 10 m.

4.5.1.2 Urban path loss model (Alternative Flat Suburban)

The COST 231 Hata model for propagation in an urban environment in the
PCS range is calculated from the expression:

Lp = 46.3 + 33.9 × Log(f ) − 13.82 × Log(hb ) − a(hm )

+ [44.9 − 6.55 × Log(hb ) ∗ Log(d) + Cm ] (4.21)

where, a(hm ) is the mobile antenna correlation factor given as:


[1.1 × Log(f ) − 0.7]hm for suburban urban
a(hm ) = −[1.56 × Log(f ) − 0.8]dB (4.22)

3.2 × [Log(11.75hm )] − 4.97 dB for dense urban
2


0 dB for suburban urban
Cm = (4.23)
3 dB for dense urban

Lp (suburban) = Lp(urban) − 2 × [Log(f/28)]2 − 5.4 (4.24)

 4.6 Walfisch–Ikegami Model 123

and
hb = base station height (m), 30–200 m
hm = mobile height (m), 1–10 m
f = frequency (MHz), 1500–2000 MHz
d = distance (km), 1–20 km
It has been shown that the COST 231 Walfish–Ikegami (W-I) model pro-
vides a close match for extensive experimental data from suburban and urban
areas and that the Category C in the Hata-Okamura model is in good agree-
ment with the cost 231 W-I model. It also provides continuity between the two
models. The COST 231 W-I model agrees well with measured data from urban
areas, provided appropriate rooftop heights and building spaces are used.

4.6 Walfisch–Ikegami Model

The Wlfisch–Ikagami Model is valid between 800 and 200 MHz and over
distances of 20 km to 5 km and was recommended by the WiMAX Forum. It
is useful for dense urban and canyon-like environments where average building
height is larger than the receiving antenna height. This means wireless signals
undergo diffraction and are guided along the street like an urban canyon. The
model applies with the following propagation parameters:
hb = base station height (m), 3–50 m
hm = mobile height (m), 1–3 m
f = frequency (MHz), 800–2000 MHz
d = distance (km), 0.2–5 km
This model applies therefore fairly well when WiMAX is deployed in
densely populated urban areas with significant number of semi-high rise build-
ings (eg. Sydney CBD). The model distinguished between the LOS and the
NLOS cases.
Line of Sight: For the line of sight case the loss equation is

L̃P = 42.6 + 26 log(d) + 20 log(f ) (4.25)

There are no correction terms in this case and it applies directly to a LOS
WiMAX situation.
For an urban canyon with line of sight the suggested propagation model is

L̃P = −31.4 + 26. log10 (d) + 20. log10 (f ) (4.26)

124 Planning of WiMAX and LTE Networks

Fig. 4.9 Communication from a MS in a Canyon.

None Line of Sight: For the none line of sight case,

L̃P = Lr + Lf + Lms (4.27)

The median path is given by the expression:

where Lr, Lf and Lms are the roof top, free space and multiscreening path
losses.

Lr + Lf + Lms ; Lr + Lms > 0
L̃P = (4.28)
Lf ; Lr + Lms ≤ 0

The free space loss is given by the expression

Lf = 32.44 + 20. log10 (f ) + 20. log10 (d) (4.29)

The frequency f is measured in MHz and the distance in km. The rooftop
diffraction model is given by the expression

Lrt = −6.9 + 10 log(w) + 10 log(f ) + 20 log(dhm ) + Lori (4.30)

Where
Lori = −10 + 0.354ϕ f or 0 ≤ ϕ < 35
= 2.5 + 0.075(ϕ − 35) f or 35 ≤ ϕ < 55 (4.30a)
= 4 − 0.114(ϕ − 55) f or 55 ≤ ϕ < 90
 4.6 Walfisch–Ikegami Model 125

Fig. 4.10 Walfisch–Ikegami Model.

L(mult) = k0 + ka + kd log(d) + kf log(f ) − 9 log(W ) (4.30b)

k0 = 0,
kd = 18 − 15(dhb /hroof )
(4.30c)
ka = 54 − 0.8dhb and
 
kf = −4 + 1.5 (f/925) − 1

For the non line of sight case in a metropolitan area with the following
parameters, the median loss model becomes


1.5f
L̃P = −65.9 + 38 log10 (d) + 24.5 + log10 (f ) (4.31)
925
126 Planning of WiMAX and LTE Networks

Table 4.7
Parameter Variable Value
Height of base station hb 12.5 m
Building height hbd 12 m
Building-to-building distance d 50 m
Building width w 25 m
Height of mobile station hm 1.5 m
Orientation of all paths ϕ 30◦

A 10 dB fading margin is suggested by the WiMAX Forum with this

formula.

4.7 Erceg Model

The Erceg model was adopted by the 802.16 group for fixed WiMAX. The
model is mostly suited to fixed wireless applications. The Erceg model consists
of a base model and an extended model. The data leading to the model were
collected in Dallas, Chicago, New Jersey, Atlanta and Seattle in the USA, at
1.9 GHz and over 95 macrocells.
The base model has three models in one and each one accounts for specified
terrain type.

i) Erceg A model is applicable to hilly terrain with moderate to heavy

density of trees
ii) Erceg B model is also applicable to hilly terrain with light tree
density. It applies also to flat terrain with moderate to heavy density
of trees
iii) Erceg C model is applicable to flat terrain with light density of trees

The Erceg loss formula models instantaneous path loss as a sum of a median
loss term and a shadow fade value given by the expression


d
LP = L̃P + X = A + 10α log10 +X (4.32)
d0
Where X is the shadow fades and A is the free space path loss over a distance
d0 = 100 m and at a frequency f.


4πf d0
A = 20 log10 (4.33)
C
 4.7 Erceg Model 127

Table 4.8 Parameters of the Erceg model.

Parameters Erceg Model A Erceg Model B Erceg Model C
a 4.6 4 3.6
b 0.0075 0.0065 0.005
c 12.6 17.1 20
Sa 0.57 0.75 0.59
µs 10.6 9.6 8.2
σs 2.3 3 1.6

The associated path loss exponent α is modelled as a Gaussian random variable

with a mean value given by the expression

A − Bhb + Ch−1
b (4.34)

Hence the instantaneous value of the path loss exponent is

α = A − Bhb + Ch−1
b + xσα (4.35)

Where x is a Gaussian random variable with zero mean and unit variance.
The standard deviation of the distribution of the path loss exponent is σα . The
parameters of the Erceg model are given below for the different terrain types.
The base model applies only at 1.9 Hz and for MS with omnidirectional
antenna at a height of 2 meters and for base stations of heights 10 to 80 m.
The extended model modifies the base model to enable it work over larger
frequency range and with the following parameters:
hb = base station height (m), 10–80 m
hm = mobile height (m), 2–10 m
f = frequency (MHz), 1900–3500 MHz
d = distance (km), 0.1–8 km
The median path loss model for the extended Erceg model is


d
L̃P = A + 10γ log10 + P .Lf + P .LhMS + P .Lθ MS (4.36)
d0
The correction terms in this equation are:


f
P .Lf = 6 log (4.37)
1900


hm
P .LhMS = −10.8 log ; for Erceg A and B (4.38)
2
128 Planning of WiMAX and LTE Networks


hm
P .LhMS = −20 log ; for Erceg C (4.39)
2



2
θ θ
P .LθMS = 0.64 ln + 0.54 ln (4.40)
360 360
The correction term P .LθMS is called the antenna gain reduction factor
because it accounts for the fact that the angular scattering is reduced due to
the directivity of the antenna. This correction can be significant and is about
7 dB at an antenna angle of 20◦ .

4.7.1 Stanford University Interim (SUI) Model

To calculate the path loss using SUI model, the environment is categorized in
three different groups with different characteristics, known as A, B, and C [1].
A is referred to the hilly environment and moderate to very dense vegetation
which results in highest path loss, while B refers to the hilly environment but
rare vegetation or high vegetation but flat terrain. However C is referred to the
flat are with rare vegetation which leads to lowest path loss.
SUI model is a suitable channel modelling for WiMAX implementation,
using frequency band at 3.5 GHz, which can support for cell radius in range
of 0.1 km and 8 km also the base station antenna height between 10 m and
80 m and receiver antenna height in the range of 2 m and 10 m [1]. In SUI
model the path loss is calculated using formula 4.41:


d
Lp = A + 10γ log10 + Xf + Xh + s f or d > d0 (4.41)
d0
In this formula, d0 = 100 m and d is the distance between the transmitter
and the receiver. s is a standard deviation which is a random variable, Xf is a
correction for frequency above 2 GHz and Xh is a correction for transmitter
antenna height [1].


f
Xf = 6.0 log10 (4.42)
2000


hr
Xh = −10.8 log10 for A and B environment (4.43)
2000


hr
Xh = −20.0 log10 for C environment (4.44)
2000
 4.8 Path Loss Exponents 129

Where f is the frequency and hr is the height of the antenna at the receiver.
However, in formula 4.41, A is calculated by:


4π d0
A = 20 log10 (4.45)
λ
Where λ is the wavelength of the signal in meter. Also γ is the path loss
exponent which has different values between 2 and 5 for different environment
and depends on the height of the base station antenna hb and three constants
of a, b and c which vary with different type of environment of A, B and C [1].
c
γ = a − bhb + (4.46)
hb
The path loss exponent in urban area when the LOS exists is 2, while it is
between 3 and 5 in the urban area with absence of a LOS. However the path
loss exponent will be more than 5 if the signal propagation is happening in an
indoor environment [1]. SUI model is used for planning for WiMAX in rural,
urban and suburban areas.

4.7.1.1 Indoor Propagation Models

In-door propagation models were originally studied for wireless LANs and
cordless phones. Propagation in buildings is highly dependent on the types
of building materials used, the lay out of the building and its contents and
location. Log-normal shadowing applies for in-door wireless applications. In
3G networks and wireless Internet access, many picocells are deployed, hence
indoor propagation modelling and estimation are of particular interest for
3G-LTE networks and WLAN.

4.8 Path Loss Exponents

Path loss exponent varies widely across propagation environments [8]. There-
fore the bound on the hop distance and number is different for different types
of propagation domains. For long-distance coverage the exponent for out door
environments is around 4 except in none line-of-sight situations when it could
be bigger than 4. The value of the path loss exponent is an indicator of how
fast energy is lost between the transmitter and receiver. α < 2 is a measure of
the guiding effect of the channel and when α > 2 the channel is considered to
be scattering energy.
130 Planning of WiMAX and LTE Networks

4.8.1 Path Loss Exponents in Different Environments

The following tables provide typical values of α and also show how different
structures guide radio waves and which ones scatter them. They also provide
a database of α for design of wireless networks in different environmental
situations.

4.8.1.1 Out Door Environments

The value of path loss exponent out doors is a function of the terrain (free
space, urban, suburban, rural and foliage type), if communication is line-
of-sight (LOS) or non-LOS (NLOS), height of the antenna and the channel
frequencies. Table 4.9 summarises these effects [8].
The dynamic range of α in this table is 6.3. None line-of-sight communi-
cation often means higher path loss exponents. Similarly the higher the height
of the antenna the higher the expected pass loss exponent. The value of the
path loss exponent after the break point is normally greater than the value

Table 4.9 Path loss exponents for out door environments.

Path Loss Exponents for Out Door Environments
No Location Path loss exponent Frequency range
1 Urban 4.2
2 Free space 2 Micro cellular
3 Log-normally shadowing area 2 to 4 Micro cellular
4 UWB LOS up to breakpoint 2 UWB range
UWB LOS after breakpoint 4 UWB range
LOS urban (antenna ht = 4 m) 1.4 5.3 GHz range
LOS urban (antenna ht = 12 m) 2.5 5.3 GHz range
NLOS urban (antenna ht = 4 m) 2.8 5.3 GHz range
NLOS urban (antenna ht = 12 m) 4.5 5.3 GHz range
5 LOS rural (antenna ht = 55 m) 3.3 5.3 GHz range
NLOS rural (antenna ht = 55 m) 5.9 5.3 GHz range
LOS suburban (antenna ht = 5 m) 2.5 5.3 GHz range
NLOS suburban (antenna ht = 12 m) 3.4 5.3 GHz range
6 Highway micro-cells 2.3 900 MHz
Dual Carriage Highway 7.7 1.7 GHz
7 BFWA/directional antenna (5.5–6.5 m) 1.6 3.5 GHz
BFWA/directional antenna (6.5–7.5 m) 2.2
BFWA/directional antenna (7.5–8.5 m) 2.7
BFWA/directional antenna (8.5–9.5 m) 2.6
BFWA/directional antenna (9.5–10.5 m) 3.6
 4.8 Path Loss Exponents 131

before the break point. The break point distance can be approximated with the
expression:
4π hT hR
db = (4.47)
λ
hT and hR are the heights of the transmitting and receiving antennas and λ is
the wavelength of transmission. The high value of exponent for dual carriage
highway is due to ground reflections from the road surface.

4.8.1.2 Indoor Environments

In Table 4.10, the path loss exponent for in door communications across a
wide variation of frequencies is shown. The unpredictability of the path loss
exponent is demonstrated by the range of values shown [8].
Communications indoors at various frequencies affect the path loss expo-
nent and the predominant sources of effects are the height of the building (or
height of antenna), antenna directivity, LOS or NLOS communication, the
channel frequencies, the types of materials used in the construction of the
buildings and the location of measurements in the building. Omni-directional
antennas often result to lower path loss exponents compared to directional
antennas. This is because, the omni-directional antennas collects signals from
many more multipath sources. Building materials of different types lead to
different path loss exponents. The dynamic range of α in this table is 8.8.
Therefore the optimum hop index will vary widely in doors. Measuring α is
therefore required prior to establishing the relay nodes.

4.8.1.3 Underground Environments

Communications underground such as in tunnels and mines forms a vital com-
ponent of the overall wireless communication industry. In many countries,
tunnels form significant sections of roads and railways. Similarly communi-
cation inside mines is also a vital support for mining and mineral exploration.
Table 4.11 records typical path loss exponents reported for underground com-
munications. Understandably, low frequency applications are prevalent.
Path loss exponent in underground communications is normally predom-
inantly very high as seen from Table 4.11 [8]. This is due to the terrain, the
materials used for construction of the tunnels and to some extent the channel
132 Planning of WiMAX and LTE Networks

Table 4.10 Path loss exponent for indoor communications.

Path Loss Exponents for Indoor Environments
No Location Path loss exponent Frequency range
1 LOS 1.83 802.11a (5.4 GHz)
2 LOS 1.91 802.11b (2.4 GHz)
3 NLOS 4.7 802.11a
4 NLOS 3.73 802.11b
5 Omni/Omnidirectional antennas 1.55 UWB
Omni/directional antennas 1.65 UWB
Directional/Directional–shadow 1.72 UWB
6 Indoor CDMA 1.8 ∼ 2.2 20 GHz–30 GHz
7 LOS 1.73 900 MHz
NLOS 0.48 ∼ 1.12 900 MHz
LOS 2.23 1.89 GHz
NLOS −1.43 ∼ 1.47 1.89 GHz
8 LOS (millimetre wave) 1.2 ∼ 1.8 94 GHz
Obstructed channel 3.6 ∼ 4.1 94 GHz
LOS 1.8 ∼ 2.0 11.5 GHz
LOS 1.2 37.2 GHz
9 Inside room of a building 0.77 900 MHz
Inside room of a building 0.44 1.35 GHz
LOS DECT picocells −1.55 1.8 GHz
NLOS DECT picocells −3.76 1.8 GHz
10 Corridor Ground Floor 0.70 450 MHz
0.48 900 MHz
0.02 1.35 GHz
−1.43 1.89 GHz
11 Corridor Floor 1 of building 1.12 450 MHz
1.02 900 MHz
0.07 1.35 GHz
1.46 1.89 GHz
12 Corridor Floor 2 of building 1.79 450 MHz
1.72 900 MHz
0.44 1.35 GHz
2.22 1.89 GHz
13 Indoor 3rd floor of a laboratory 1.3 2.45 GHz
1.8 5.25 GHz
1.7 10 GHz
1.8 17 GHz
1.7 24 GHz

frequencies used. The scattering properties of the terrain also affect the path
loss exponent. The dynamic range of α in this table is 10.95. Path losses under-
ground are therefore very high and hop distances must be chosen with this in
mind.
 4.8 Path Loss Exponents 133

Table 4.11 Path loss exponent for underground communications.

Path Loss Exponents for Underground Communications
No Location Path loss exponent Frequency range
1 Underground (train)–front 12.45 465 MHz
2 Underground (train)–rear 9.72
3 Underground (train)–front 8.58 820 MHz
4 Underground (train)–rear 8.17
5 Train yard (parallel to track) 2.7
Train yard (cross-track) 3.4
6 Underground Mine 2.13 ∼ 2.33 2.4 GHz
Moving train–140 km track sites 1.5 ∼ 7.7 320 MHz

Table 4.12 Effect of materials on path loss exponent.

Path Loss Exponents for Different Environmental Structures
No Location Path loss exponent Frequency range
1 Engineering 1.4 ∼ 2.2 0.8 GHz–1.0 GHz
2 Apartment Hallway 1.9 ∼ 2.2
3 Parking structure 2.7 ∼ 3.4
4 One-sided corridor 1.4 ∼ 2.4
5 One-sided Patio 2.8 ∼ 3.8
6 Concrete Canyon 2.1 ∼ 3.0
7 Plant fence 4.6 ∼ 5.1
8 Small Boulders 3.3 ∼ 3.7
9 Sandy Flat beach 3.8 ∼ 4.6
10 Dense Bamboo 4.5 ∼ 5.4
11 Dry Tall Underbrush 3.0 ∼ 3.9

Table 4.13 SNR parameters as function of modulation schemes.

SNR CC (AWGN, SNR CTC (AWGN, Data bit per
Modulation scheme BER 10–6) BER 10-6) symbol
QPSK1/2 5 dB 2.5 dB 1
QPSK 3/4 8 dB 6.3 dB 1.5
16-QAM 1/2 10.5 dB 8.6 dB 2
16-QAM 3/4 14 dB 12.7 dB 3
64-QAM 1/2 16 dB 13.8 dB 3
64-QAM 2/3 18 dB 16.9 dB 4
64-QAM 3/4 20 dB 18 dB 4.5

4.8.1.4 Unspecified Environments

Path loss exponent in other terrains of interest are summarised in Table 4.12 [8].
Table 4.12 demonstrates the varying nature of path loss exponent when
different types of materials and the terrain types that affect communications
are considered. These tables show that there is no universally accepted path
 Table 4.14 WiMAX link budget [2].
Mobile Handheld in Outdoor Scenario Fixed Desktop in Indoor Scenario
Parameter Downlink Uplink Downlink Uplink Notes
Power amplifier output power 43.0 dB 27.0 dB 43.0 dB 27.0 dB AI
Number of tx antennas 2.0 1.0 2.0 1.0 A2
Power amplifier backoff 0 dB 0 dB 0 dB 0 dB A3: assumes that amplifier has sufficient
linearity for QPSK operation without
backoff
Transmit antenna gain 18 dBi 0 dBi 18 dBi 6 dBi A4: assumes 6 dBi antenna for desktop SS
Transmitter losses 3.0 dB 0 dB 3.0 dB 0 dB A5
Effective isotropic radiated power 61 dBm 27 dBm 61 dBm 33 dBm A6 =
A1 + 10 log10 (A2) − A3 − +A4 − A5
Channel bandwidth 10 MHz 10 MHz 10 MHz 10 MHz A7
Number of subchannels 16 16 16 16 A8
134 Planning of WiMAX and LTE Networks

Receiver noise level −104 dBm −104 dBm −104 dBm −104 dBm A9 = −174 + 10 log10 (A7∗ le6)
Receiver noise level 8 dB 4 dB 8 dB 4 dB A 10
Required SNR 0.8 dB 1.8 dB 0.8 dB 1.8 dB A11: for QPSK, R 1/2 at 10% BLER in ITU
Ped. B channel
Macro diversity gain 0 dB 0 dB 0 dB 0 dB A12; No macro diversity assumed
Subchannelization gain 0 dB 12 dB 0 dB 12 dB A13 = 10 log10 (A8)
Data, rate per subchannel (kbps) 151.2 34.6 151.2 34.6 A14; using QPSK, R 1/2 at 10% BLER
Receiver sensitivity (dBm) −95.2 −110.2 −95.2 −110.2 A15 = A9 + A10 + A11 + A12 − A13
Receiver antenna gain 0 dBi 18 dBi 6 dBi 18 dBi A16
System gain 156.2 dB 155.2 dBi 162.2 dB 161.2 dB A17 = A6 − A15 + A16
Shadow-fade margin 10 dB 10 dB 10 dB 10 dB A18
Building penetration loss 0 dB 0 dB 10 dB 10 dB A19; assumes single wall
Link, margin 146.2 dB 145.2 dB 142.2 dB 141.2 dB A20 = A17 − A18 − A19
Coverage range 1.06 km (0.66 miles) 0.81 km (0.51 miles) Assuming COST-231 Hata urban model
Coverage range 1.29 km (0.80 miles) 0.99 km (0.62 miles) Assuming the suburban model
 References 135

loss model for indoor, out door or underground channels. The path loss model
varies, from building to building and from terrain to terrain. All path loss
models in use are approximations for only a few conditions.

4.8.1.5 Receiver Sensitivity

The receiver power is limited by its design (implementation margin), thermal
noise in the receiver, its noise figure and signal-to-noise ratio. Thus the receiver
sensitivity is given by the expression:

RSS = SNRR + N FR + Limplementation + NThermal (4.48)

These factors are given by the expressions


NUsed
NThermal = −147 + 10 log10 (f ) = −147 + 10 log10 B.n. (4.49)
NFFT
Where f is the subcarrier spacing. The thermal noise is affected by its
thermal noise density which is

N0 = K.T .B ≈ −174 dB (4.50)

K is Boltzman constant. The SNR is a function of the modulation scheme used

and these have been provided in the standard and are given in Table 4.13.
The receiver noise figure is caused by the electronics in its RF chain. It
is a ratio of the input SNR to a device to its output SNR. This is normally
measured at 290 Kelvin. It provides a measure of the performance of a device.
It is given by the expression:
SNRin
NF = (4.51)
SNRout
Table 4.14 is a summary of a WiMAX link budget [2].

References
[1] Josip Milanovic, Snjezana Rimac-Drlje, Krunoslav Bejuk, “Comparison of Propaga-
tion Models Accuracy for WiMAX on 3.5 GHz.” IEEE International Conference,
pp. 111–114, 2007.
[2] Mobile WiMAX Group, “Coverage of mobile WiMAX”, pp. 1-18.
[3] Rana Ezzine, Ala Al Fuqaha, Rafik Braham, Abdelfettah Belghith. “A New Generic
Model for Signal Propagation in WiFi and WiMAX Environment”. Wireless Days,
pp. 1–5, 2008.
 136 Planning of WiMAX and LTE Networks

[4] Tamaz Javornik, Gorazd Kandus, Andrej Hrovat, Igor Ozimek. Software in Telecommu-
nication and Computer Networks. pp 71–75, 2006.
[5] Yan Zhang. WiMAX Network Planning and Optimization. USA: CRC Press, 2009.
[6] Johnson I Agbinya. IP Communications and Services for NGN. New York: Taylor and
Francis, 2009.
[7] Kejie Lu, Yi Qian, Hsiao-Hwa Chen, Shengli Fu. “ WiMAx Networks: From Access to
Service Platform”. IEEE Computer Society, 22, pp. 38–45, May/June 2008.
[8] Johnson I Agbinya, “Design Consideration of Mohots and Wireless Chain Networks”,
Wireless Personal Communication”, © Springer 2006, Vol. 40, pp. 91–106.

View publication stats