(IJCSIS) International Journal of Computer Science and Information Security, 2009

Vol. 4, No. 1 & 2, 2009

Channel Equalization in Digital Transmission

Kazi Mohammed. Saidul. Huq #1, Miguel Bergano#1,Atilio Gameiro #1, Md. Taslim Arefin *2

# *
Institute of Telecommunications Lecturer, Dept. of Computer Science and Engineering
Aveiro, Portugal University of Development Alternative (UODA)
[email protected]
Dhanmondi, Dhaka-1209, Bangladesh.
2
[email protected]

Abstract— Channel equalization is the process of reducing

amplitude, frequency and phase distortion in a radio channel
with the intent of improving transmission performance. Different
types of equalizers, their applications and some practical
example is given. Especially, in the digital communication
scenario how equalization works is shown. This paper presents a
vivid description on channel equalization in digital transmission Figure 1: Elements of a communication System
system.
In a real communication system the communication channel is
Keywords— ISI, Baseband, Passband, equalization not perfect, perturbations caused by imperfections on the
transmission channel or interference from outside world, for
1. INTRODUCTION example, can generate a bad functionality of the channel.
Having these issues the channel will not perform a flat
A communication system is basically a way of transmitting frequency response and linear phase shift mainly because of
information trough a communication channel and usually distortion. Interference and noise are contaminations that
associated with it are the transmitter and a receiver. The main occur from other radio systems and from random electrical
function of it is to guarantee that information, or message, signals produced by natural processes, respectively. In order
from the transmitter should be available at the receiver to perform a good way of conveying information from
without perturbations. A communication system is completed transmitter to receive the problems mentioned earlier should
when joining these three parts, the transmitter, the receiver be considered in modeling a communication system. The main
and the communication channel. Examples of communication task in this procedure is to take the channel conditions and in
channels are telephone channels, coaxial cables, optical fiber, some way invert it, or in other words, a channel can be
wireless broadcast channel, mobile radio channel and satellite mathematically estimated by a transfer function, at the output,
channels. The signal to be transmitted could be analog or or at the receiver, it would be a system with an inverse of that
digital. The first one implies the use of fewer hardware on the transfer function. Some problems arise in modeling the
receiver and transmitter, on the contrary digital signals need channel; issues like nonlinearity or time variance induce
more hardware, although digital systems are more stable, difficulties. All these mentioned issues are obstacles to
flexible and more reliable. It should be noted, however, that approach an ideal frequency response of the communication
we can implement much of an analog communication system system or to identify the channel characteristics exactly.
using digital hardware and the appropriate ADC and DAC
steps, and thereby secure an analog system many of the 1.1. Digital Transmission
advantages of a digital system.
A digital transmission performs digital messages that are
Ideally a system like this would work perfectly but due to basically ordered sequence of symbols produced by a discrete
imperfections of the channel it can be defined by a more information source. Here the task is to transfer a digital
complete diagram, represented in Fig. 1. message from the source to the destination. In an analog
communication system problems like the channel frequency
bandwidth and the signal to noise ratio cause errors that
appear in the received message, similarly, signaling rate and

ISSN 1947 5500

 (IJCSIS) International Journal of Computer Science and Information Security, 2009
Vol. 4, No. 1 & 2, 2009

error probability play roles in digital communication namely

in the output messages.

Digital signal has usually a form of an Amplitude Modulated

Pulse Train and are commonly expressed by:
x(t) = ∑ ak p (t − kD) (1)
k
where ak represents the modulated amplitude of each symbol k,
D is for pulse duration or pulse to pulse interval and p(t) is the
unmodulated pulse that has the values 1 or 0 periodically. In
the case of binary signaling D indicates bit duration so D=Tb,
and the bit rate is rb=1/Tb measured in symbols per second or
baud. Digital PAM signals can take several formats, basically
a simple on off with a defined duration generates a format
called RZ (return to zero), but others exist like NRZ (Non
return to Zero), both polar having a DC component that
Figure 2: Binary PAM formats
wastes power, an alternative is bipolar NRZ or split-phase
Manchester or even quaternary. In a transmission process
there are several perturbations, noise contamination cross talk Always associated in all these formats are the noise reduction
or spill over from other signals a phenomena described as ISI by introducing a filter, this filter should not introduce ISI, like
– Inter Symbol Interference – that basically is a form of showed in Fig. 3.
distortion of signals in which symbols interfere with
subsequent symbols. Reducing the bandwidth of filter will
reduce noise but would increase the ISI, for that Nyquist
stated that the symbol rate r must be lower than the twice of
channel bandwidth.
r ≤ 2B (2)
On the list of the digital transmission limitations is obvious Figure 3: Baseband transmission system
the channel, so to approach an ideal frequency response the The amplifier compensates losses in the channel and the filter
channel must be equalized. The equalizer is usually inserted LPF removes out of band contaminations, the output message
between the receiver and the channel regenerator. With this, it is the recovered message from the digital signal.
will increase the knowledge of the channel characteristics that
sometimes results in some residual ISI. An equalizer is based To transmit in longer distances passband digital transmission
on the structure of a transversal filter, like it will be shown is used, and requires modulation methods applied in analog
later. signals. Digital information has a lot of ways to be performed
in a carrier wave, it can modulate amplitude, frequency or
1.2. Baseband and passband digital transmission phase of a sinusoidal carrier waive.

At baseband a digital message is represented by a PAM and Any modulated passband signal may be expressed in the
expressed like equation (1). Above the modulated forms that a quadrature-carrier form:
baseband signal can take was already mentioned, RZ,
NRZ(NRZ-L, NRZ-M, NRZ-S), Bipolar, Biphase (Biphase-L, x c (t) = Ac [x i (t)cos(wc t + θ ) − x q (t)sin(wc t + θ )] (3)
Biphase-M, Biphase-S), Differential Manchester:
The carrier frequency fc, amplitude Ac and phase are constant.
The message is contained in the phase – i – and quadrature – q
– components. An amplitude modulation (ASK – Amplitude
Shift Keying) can be achieved simply using a NRZ signal,
another example is QAM (Quadrature Amplitude Modulation)
that achieves higher modulation speed. Phase Shits can also
perform phase modulation often described as BPSK (Binary
Phase Shift Keying, if the signal has four elements in the
alphabet the modulation is QPSK (Quaternary Phase Shift
Keying). An example of transmitter is in Fig. 4:

ISSN 1947 5500

 (IJCSIS) International Journal of Computer Science and Information Security, 2009
Vol. 4, No. 1 & 2, 2009

x(t) is for the output of the filter. At the receiver is the filter,
with frequency response HR(t) or impulse response hR(t), a
sampler and a comparator. So at the ouput it will be:
+∞
v(t) = y(t) ∗ hR (t) = ∑ Aa p (t − t
k r d − KTb ) + n 0 (t) (7)
k=−∞
The value A is a scale factor such that pr(0)=1, n0(t) is the
noise component at the receiver filter output, and pr(t-td) is the
pulse shape at the receiver filter output, delayed by an amount
td due to filtering.
Having all the information of the response of a
communication system it is possible to develop forms to
Figure 4: QPSK Transmitter minimize problems in the system, like ISI and SNR reduction.
Zero ISI and and noise can be achieved by choosing the
correct HT(f) and HR(f). The equations, given in [1][2],
A frequency modulation (FSK – Frequency Shift Keying) is demonstrate that is a hard task to create such frequency
obtained when the input signal x(t) selects the frequency of an responses mainly because of the channel conditions, in a
oscillator. baseband transmission, or PAM, like a modem, it must have
information about the channel. For passband transmission,
2. PRINCIPLES like cellular radio there are several obstacles to the
transmission, or for microwave links that depend on the
Previously was defined several concepts of digital atmosphere conditions. So the best filter to use at the receiver
transmission systems, namely its limitations (bandwidth, noise, must be adjustable improving the performance of a
distortion and ISI), and formats of transmitting (modulation) transmission. Such filter is called equalizer. There are two
for baseband and passband transmission. In order to avoid the types of equalizers: preset and adjustable. The first one its
issues related to these types of communications systems it parameters are determined by making measurements on the
must properly designed. channel and determining these parameters using these
measurements. The adaptive, is automatic, its parameters are
adjusted by sending a known signal, called training signal.

Transmitter Channel
Source Equalizer
Filter Filter
Figure 5: Ideal model of a communication System
Figure 6: Block Diagram of PAM Communication System with
The signal source input has the regular input represented by (1) equalization
this time p(t) has the form of a unit pulse δ(t). The next The previous figure illustrates the process of equalization. The
subsystem is transmission filter with low pass frequency overall frequency response is:
response HT(f) or impulse response hT(t). The transmitted
signal is given by:
+∞ +∞
H ( f ) = H ( f )H ( f )H ( f ) (8)
0 T C E

x t (t ) = ∑ a δ(t − kT )∗ h
k b T (t) = ∑ a h (t − kT )
k T b
(4)
In theory an equalizer should have an impulse response that is
k=−∞ k=−∞
the inverse of that on the channel, and the design of this
Where the asterisk represents convolution. The channel systems involves a compromise between ISI reduction and
can be considered as a filter, due to its bandwidth limitations noise reduction of the channel.
and imposes a frequency response function HC(f) or impulse
response hC(t) and additive Gaussian Noise represented by n(t).
At the receiver will be: 3. TYPES OF EQUALIZERS

y(t) = x(t) + n(t) (5) 3.1. Zero forcing

x(t) = x t (t) ∗ hC (t) (6)
The basic idea of a Zero-Forcing Equalization – ZFE – is to
implement a filter (equalizer) that follows the channel
response, or like already said, the channel filter. The system of

ISSN 1947 5500

 (IJCSIS) International Journal of Computer Science and Information Security, 2009
Vol. 4, No. 1 & 2, 2009

a ZFE has a frequency response indicated in (8). Asuming that MMSE = E [(error) 2 ] (9)
the first Nyquist critereon is satisfied by the sampler a ZFE is
like a inverse filter, the inverse frequency response of the Analitically the error represents the difference between the
chanell frequency response and is usually approximated by a desired value and the real value.
set of FIR filters like is presented in Fig. 7.

To formulate a set of FIR inverse filter coefficients, a training

{
MMSE = E [z( t ) − d (t )]
2
} (10)
signal consisting of an impulse is transmitted over the channel. Following this concept of obtaining the minimum error, the
By solving a set of simultaneous equations based on the task is to determine the taps of the filter in Fig. 7 in order to
received sample values, a set of coefficients can be perform a transmission with minimum errors. In Fig. 8 is
determined to force all but the center tap of the filtered presented a scheme points out the interesting signals used in
response to 0. This means the N–1 samples surrounding the the process[10].
center tap will not contribute ISI. The main advantage of this y(t)
Transverse z(t)
technique is that the solution to the set of equations is reduced Channel and
Receiver ∑ Filter
Decision
to a simple matrix inversion. Filters
Equalizer

The major drawback of ZFE is that the channel response may Noise ∑
often exhibit attenuation at high frequencies around one-half d’(t)
the sampling rate (the folding frequency). Since the ZFE is Transverse
simply an inverse filter, it applies high gain to these upper Filter
Equalizer
frequencies, which tends to exaggerate noise. A second
problem is that the training signal, an impulse, is inherently a
low-energy signal, which results in a much lower received
signal-to-noise ratio than could be provided by other training
signal types [3][6]. Figure 8: MMSE Equalizer Circuit

3.3. Adaptive equalizers

Most of the times the channel, besides being unknown, it is

also changing with time, a solution can be achieved by
creating an algorithm that adjust the taps of the filter by
following the channel and lead to the optimum values of the
equalizer. Adaptive equalization has different ways to perform
automatic algorithms.

3.3.1 Decision Directed Equalization

Figure 7: Filter Structure of a ZFE

The previous equalizer systems are linear in that they employ
linear transversal filter structures. The filters implement a
3.2. Minimum Mean Square convolution sum of a computed impulse response with the
input sequence. Often with data communication systems, one
Since ZFE ignore the additive noise and may significantly can take advantage of prior knowledge of the transmit signal
amplify noise for channels with spectral nulls another type of characteristics to deduce a more accurate representation of the
equalizer my be used to partially avoid this problem. The transmit signal than can be afforded by the linear filter. It is
Minimum-mean-square error (MMSE) equalizers minimize possible to devise a decision device (a predictor or a slicer)
the mean-square error between the output of the equalizer and that estimates what symbol value was most likely transmitted,
the transmitted symbol. They require knowledge of some auto based on the linear filter continuous output. The difference
and cross-correlation functions, which in practice can be between the decision device input and output forms an error
estimated by transmitting a known signal over the channel. In term which can then be minimized to adapt the filter
such an equalizer the Coefficients in Fig. 7 are chose to coefficients. This is true because a perfectly adapted filter
minimize the mean square error (MMSE). The error consists would produce the actual transmitted symbol values, and,
on the sum of the squares of ISI terms plus noise. therefore, the slicer error term would go to 0. In practice, the
The MMSE at the equalizer is the expected value of the error is never 0, but if the adapted filter is near ideal, the
square of the error.

ISSN 1947 5500

 (IJCSIS) International Journal of Computer Science and Information Security, 2009
Vol. 4, No. 1 & 2, 2009

decisions are perfect. In this case, the slicer is effectively channel filter are calculated. It verifies a good estimation of
throwing away received noise with each decision made. the channel filter parameters as demonstrated by the error
curve, that present values of 10-1. Also it verifies very
3.3.2 Decision-Feedback Equalization approximate values of the weights.

Another nonlinear adaptive equalizer should be considered:

the decision feedback equalization (DFE). DFE is based on
the principle that once we have determined the value of the
current transmitted symbol, we can exactly remove the ISI
contribution of that symbol to future received symbols (see
Figure 5). The nonlinear feature is again due to the decision
device, which attempts to determine which symbol of a set of
discrete levels was actually transmitted. Once the current
symbol has been decided, the filter structure can calculate the
ISI effect it would tend to have on subsequent received
symbols and compensate the input to the decision device for
the next samples. This postcursor ISI removal is accomplished
by the use of a feedback filter structure.
Figure 9: Real and Estimated Output Signals

4. DESIGN IN BASEBAND

In baseband the frequency bandwidth of transmission is equal

to the symbol rate. In this case the samples are real numbers
while passband the samples are complex numbers. A first
consequence of baseband equalization is the delay
introduced by the equalizer in the carrier recovery loop. This
delay affects the loop stability, steady-state jitter
performance as well as its acquisition behavior. An example
of a Least Mean Square Algorithm is presented next.

4.1. Program in Matlab

The least mean squared (LMS) equalizer is a more general Figure 10: Error
approach to automatic synthesis. The coefficients are
gradually adjusted to converge to a filter that minimizes the
error between the equalized signal and the stored reference.
The filter convergence is based on approximations to a
gradient calculation of the quadratic equation representing the
mean square error. The only parameter to be adjusted is the
adaptation step size αa. Through an iterative process, all filter
tap weights are adjusted during each sample period in the
training sequence. Eventually, the filter will reach a
configuration that minimizes the mean square error between
the equalized signal and the stored reference. As might be
expected, the choice of αa involves a tradeoff between rapid
convergence and residual steady-state error. A too-large
setting for αa can result in a system that converges rapidly on
start-up, but then chops around the optimal coefficient settings Figure 11: Estimated and Calculated Weights of Channel Filter
at steady state.
In this algorithm the input signal considered was noise, and A MMSE algorithm was also tested and presented a clear way
the channel filter parameters were previously determined in a of the implementation of this type of equalizer. The plots
practical experience. Noise (White Noise) was added to the show the equalizer results for 1000 samples and using 500 for
output of the channel. Then it takes N samples for training training. The input is a QAM signal.
sequence (for the plots N=60) and finally the weights of the

ISSN 1947 5500

 (IJCSIS) International Journal of Computer Science and Information Security, 2009
Vol. 4, No. 1 & 2, 2009

5. BASEBAND VS PASSBAND EQUALIZATION

5.1. Examples

Next is an example of a Baseband and PassBand Equalization

in the context of QAM (or multiphase PSK).[8]
In a practical implementation, an equalizer can be realized
either at baseband or at passband. For example Fig. 15
illustrates the demodulation the demodulation of QAM (or
multiphase PSK) by first translating the signal to baseband
and equalizing the baseband signal with an equalizer having
complex-valued coefficients.

cos ωi t
Figure 12: Input Signal
Re ⎡ˆIn ⎤
⎣ ⎦

Im ⎡ˆIn ⎤
⎣ ⎦

Re [ε n ]

Im [ε n ]

sin ωi t

Figure 15: QAM (Multiphase QPSK) signal demodulation

In effect, the complex equalizer with complex-valued (in-

phase and quadrature components) inputs is equivalent to four
parallel equalizers with real-valued tap coefficients as shown
in Fig. 16.
Figure 13: Received Samples

[ Re(cn )]
Re ⎡ ˆI n ⎤
⎣ ⎦

[ Im(cn )]

Im ⎡ ˆI n ⎤
[ Im(cn )] ⎣ ⎦

[ Re(cn )]

Figure 16: Complex-valued baseband equalizer for QAM (Multiphase

QPSK) signals

On the other hand, we may equalize the signal at passband.

Figure 14: Equalized Symbols This is accomplished as shown in Fig. 17 for a two-
It performs a good estimation of the weights of the dimensional signal constellation such as QAM and PSK. The
transversal filter, and provides the optimum values for the received signal is filtered and, in parallel, it is passed through
filter. a Hilbert transformer, called a phase-splitting filter.

ISSN 1947 5500

 (IJCSIS) International Journal of Computer Science and Information Security, 2009
Vol. 4, No. 1 & 2, 2009

e − jω t system to implement in a consumer system. As discussed

below, there are other designs that outperform the DFE in
terms of convergence or noise performance, but these
generally come at the expense of greatly increased system
complexity. Today, most TDMA phones employ DFE running
on fixed-point DSPs such as those in the TMS320C5x [4]
family.

6.3. Equalization used in GSM

e jω t

Figure 17 – QAM or QPSK signal equalization at passband signals An adaptive equalizer is used in the demodulator of the
receiver to compensate for its difficulty in recognizing the
Thus, we have the equivalent of in-phase and quadrature
original bit pattern from the distorted signal. Distortion of the
components at passband, which are fed to a passband complex
signal is caused by the fact that the Doppler shift and the delay
equalizer. Following the equalization, the signal is down-
time for each path varies continuously. As a result, the
converted to a baseband and detected. The error sigbal
channel characteristic (the impulse response) changes over
generated for the purpose of adjusting the equalizer co-
time. The equalizer used for GSM is specified to equalize
efficients is formed at baseband and frequency-translated to
echos up to 16 ms after the first signal received. This
passband as illustrated in Fig. 17.
corresponds to 4.8 km in distance. One bit period is 3.69 ms.
Hence, echos with about 4 bit lengths delay can be
6. APPLICATIONS (EXAMPLES)
compensated [5].
6.1. Equalization in Modem (ADSL) Applications
6.4. Equalization in HSPA and 3GPP
Today, automatic equalization is used on just about all
Receiver-side equalization [6] has for many years been used
modems designed for operation over the switched telephone
to counteract signal corruption due to radio-channel frequency
network. With automatic equalization, a certain initialization
selectivity. Equalization has been shown to provide
time is required to adapt the modem to existing line conditions.
satisfactory performance with reasonable complexity at least
This initialization time becomes important during and after
up to bandwidths corresponding to the WCDMA bandwidth
line outages, since line initial equalization times can extend
of 5MHz [7]. However, if the transmission bandwidth is
otherwise short dropouts unnecessarily. Recent modem
further increased up to, for example 20 MHz, which is the
developments shortened the initial equalization time to
target for the 3GPP Long-Term Evolution, the complexity of
between 15 and 25 ms, whereas only a few years ago a much
straightforward high-performance equalization starts to
longer time was commonly required. After the initial
become a serious issue. One option is then to apply less
equalization, the modem continuously monitors and
optimal equalization, with a corresponding negative impact on
compensates for changing line conditions by an adaptive
the equalizer capability to counteract the signal corruption due
process. This process allows the equalizer to ‘track’ the
to radio-channel frequency selectivity and thus a
frequently occurring line variations that occur during data
corresponding negative impact on the radio-link performance.
transmission without interrupting the traffic flow. On one
The use of specific single-carrier transmission schemes,
9600 bps modem, this adaptive process occurs 2400 times a
especially designed to allow for efficient but still reasonably
second, permitting the recognition of variations as they
low-complexity equalization.
occur[9].
Linear time-domain (frequency-domain) filtering/equalization
implies that linear processing is applied to signals received at
6.2. Equalization for Digital Cellular Telephony
different time instances (different frequencies) with a target to
maximize the post-equalizer SNR (MRC-based equalization),
The direct sequence spreading employed by CDMA (IS-95)
alternatively to suppress signal corruption due to radio-
obviates the need for a traditional equalizer. The TDMA
channel frequency selectivity (zero-forcing equalization,
systems (for example, GSM and IS-54), on the other hand,
MMSE equalization, etc.).
make great use of equalization to contend with the effects of
multipath-induced fading, ISI due to channel spreading,
7. CONCLUSION
additive received noise, and channel-induced spectral
distortion, etc. Because the RF channel often exhibits spectral
nulls, the linear equalizers are not optimal due to their Of particular interest today is the area of digital cellular
tendency to boost noise at the null frequencies. Of the communications, which has seen wide use of fixed-point
nonlinear equalizers, the DFE is currently the most practical DSPs. DSP-based equalizer systems have become ubiquitous

ISSN 1947 5500

 (IJCSIS) International Journal of Computer Science and Information Security, 2009
Vol. 4, No. 1 & 2, 2009

in many diverse applications including voice, data, and video [9]. Peebles, P.Z., Communication System Principles,
communications via various transmission media. Typical Addison-Wesley, 1976.
applications range from acoustic echo cancelers for full- [10]. Samueli, H., Daneshrad, B., Joshi, R., Wong, B., and
duplex speakerphones to video echo-canceling systems for Nicholas, H., “A 64-Tap CMOS Echo
terrestrial television broadcasts to signal conditioners for Canceller/Decision Feedback Equalizer for 2B1Q
wireline modems and wireless telephony. HDSL Transceivers”, IEEE Journal onSelected Areas
in Communications, Vol. 9, Iss: 6 , August 1991, pp.
The effect of an equalization system is to compensate for 839–847.
transmission-channel impairments such as frequency-
dependent phase and amplitude distortion. Besides correcting
for channel frequency-response anomalies, the equalizer can
cancel the effects of multipath signal components, which can AUTHORS PROFILE
manifest themselves in the form of voice echoes, video ghosts
or Rayleigh fading conditions in mobile communications
channels. Equalizers specifically designed for multipath Kazi Mohammed Saidul Huq received B.Sc. in CSE from
correction are often termed echo-cancelers. They may require Ahsanullah University of Science & Technology, Bangladesh
significantly longer filter spans than simple spectral equalizers, in 2003. He obtained his M.Sc. in EE - specialization
but the principles of operation are essentially the same. Telecommunications from Blekinge Institute of Technology,
Sweden in 2006. Since April 2008, he started working at
This article attempts to familiarize you with some basic Instituto de Telecomunicações, Pólo de Aveiro, Portugal. His
concepts associated with channel equalization and data research activities include integration of heterogeneous
communication in general. This report is intended to give an wireless systems (in CRRM, cross-layer design, DBWS &
introduction to equalization, their types and examples and system level simulation paradigm) and integration of RFID.
applications in digital transmission. We have provided a brief
survey of equalization techniques and describe their Atílio Gameiro received his Licenciatura (five years course)
characteristics using some examples. Baseband and Passband and his PhD from the University of Aveiro in 1985 and 1993
equalization has been discussed in terms of Multiphase QPSK. respectively. He is currently a Professor in the Department of
Some Matlab driven examples also shown using plot to better Electronics and Telecommunications of the University of
understand. Aveiro, and a researcher at the Instituto de Telecomunicações
- Pólo de Aveiro, where he is head of group. His main
interests lie in signal processing techniques for digital
REFERENCES communications and communication protocols. Within this
research line he has done work for optical and mobile
[1]. B. P. Lathi, Modern Digital and Analog communications, either at the theoretical and experimental
Communication Systems, Third Edition: Oxford level, and has published over 100 technical papers in
University Press, 1998. International Journals and conferences. His current research
[2]. Ziemer, R.E., and Peterson, R.L., Introduction to activities involve space-time-frequency algorithms for the
Digital Communication, Second Edition, Prentice Hall, broadband component of 4G systems and joint design of
2001. layers 1 and 2.
[3]. J. Kurzweil, An Introduction to Digital
Communications, John Wiley, 2000. Md. Taslim Arefin received B.Sc. in Computer Engineering
[4]. TMS320C5x User’s Guide, Texas Instruments, 1993. from American International University –Bangladesh (AIUB)
[5]. GSM Introduction WL9001student guide Lucent in 2005. He obtained his M.Sc. in Electrical Engineering –
Technologies, 1998. Specialization Telecommunications from Blekinge Institute of
[6]. J.G. Proakis, Digital Communications, McGraw-Hill, Technology (BTH), Sweden in 2008 . At the present time he
New York, 2001. is working as lecturer in the Dept. of Computer Science &
[7]. G. Bottomley, T. Ottosson and Y.-P. Eric Wang, ‘A Engineering at University of Development Alternative
Generalized RAKE Receiver for Interference (UODA), Dhaka, Bangladesh from January, 2009. His
Suppression’, IEEE Journal on Selected Areas in research interest includes BSS, communication engineering
Communications, Vol. 18, No. 8, August 2000, pp. and computer networking like development over cellular
1536–1545. network, routing related issue and wireless communication
[8]. Qureshi, S.,“Adaptive Equalization”, IEEE etc.
Communications Magazine, March 1992, pp. 9–16.

ISSN 1947 5500