Reem ICI Self Cancellation Book Chapter BP International

Chapter 9
Print ISBN: 978-93-5547-370-7, eBook ISBN: 978-93-5547-378-3
An Efficient ICI Self-cancellation Scheme for High

order Constellations
S. E. D. Habib a*, Reem I. Sayed b, Hisham M. Hamed c and Magdi Fikri a
DOI: 10.9734/bpi/tier/v1/2265B
ABSTRACT
OFDM is an established modulation scheme in current broadband wireless mobile communication

systems due to its high spectral efficiency and robustness to multipath interference. A well-known
disadvantage of the OFDM system is its sensitivity to frequency offset caused by oscillator
inaccuracies and the Doppler shift in the channel. The carrier frequency offset (CFO) causes loss of
orthogonality, in which case the signals transmitted on each carrier are not independent of each other,
thus leading to Inter Carrier Interference (ICI).
Several techniques were developed to reduce the effect of frequency offset. These techniques can be
grouped into two broad categories. The first category is termed the frequency offset estimation and
correction category. The second group is termed the frequency-offset reduction category. In this
second category, redundant modulation is necessary in order to achieve a trade-off between
frequency offset sensitivity and bandwidth efficiency.
In this chapter, a new scheme for ICI self-cancellation is proposed. The main idea of this scheme is
first introduced using a simple single carrier system, and then it is applied to multicarrier modulation
systems. In the proposed scheme each symbol is conjugated and repeated to build a data pattern in
the transmitted stream. This data pattern will be utilized at the receiver to cancel the CFO errors.
Thus, the transmitted stream takes the form X1X1*X2X2*X3X3* …. .At the receiver, each symbol
suffers a CFO phase error. By calculating the phases of the received symbols R1A and R1B*
corresponding to transmitted symbols X1 and X1*, averaging these two phases, the corrected
received signal is calculated. Our proposed ICI self-cancellation scheme is, therefore, labeled as
Conjugate Phase Additive (CPA) scheme. We apply the proposed single carrier modulation scheme
to the OFDM system, and we propose two new conjugate techniques; Adjacent Conjugate Phase
Additive OFDM (ACPA) and Symmetric Conjugate Phase Additive OFDM (SCPA). The proposed
schemes are implemented and simulated in MATLAB environment. Their results are compared to the
existing ICI self-cancellation schemes. In contrast to previous ICI cancellation schemes, these two
schemes display a more graceful degradation of Bit Error Rate (BER) with the increase of the
normalized frequency offset. Compared to the best published ICI self-cancellation scheme (ASR,
SCSR), the SCPA scheme yields 7 dB improvement in BER over that of the SCSR at a normalized
frequency offset = 0.1 for 256-QAM modulation type.
The proposed 256-QAM ACPA OFDM ICI self-cancellation system is implemented using Xilinx
System Generator tool for DSP design, co-simulated on Spartan3A DSP 3400 device FPGA, and
compared with widely used CFO Error estimation and Correction technique (CFO-EC), using cyclic
prefix, in terms of hardware resource usage and frequency offset analysis. The CFO correction
hardware required for the ACPA scheme is less than 50% of that required for the classical CFO-EC
scheme. Relative to the classical 128-QAM OFDM system with CFO-EC correction, the 256-QAM
ACAP scheme achieves comparable immunity against frequency offset errors over the practical
frequency offset range.
Keywords: OFDM; self-cancellation system; ICI; Conjugate Phase Additive OFDM; CPA
_____________________________________________________________________________________________________
a
Department of Electronics and Electrical Communications Engineering, Cairo University, Cairo, Egypt.
b
Electronics Factory, AOI, Egypt.
c
Wasiela Company, Egypt.
*Corresponding author: E-mail: seraged11@gmail.com;
Technological Innovation in Engineering Research Vol. 1
An Efficient ICI Self-cancellation Scheme for High order Constellations
1. INTRODUCTION
Orthogonal Frequency Division Multiplexing (OFDM) is commonly considered to be the technical

solution for fourth Generation Cellular Systems (4G), also known as Long Term Evolution (LTE).
OFDM is robust in fast fading channel conditions and allows a high level of spectral efficiency.
Multiple access techniques which are quite developed for the single carrier modulations (e.g. TDMA,
FDMA) had made it possible for one communication medium to be shared by multiple users
simultaneously [1,2]. The sharing is required to achieve high capacity by simultaneously allocating the
available bandwidth to multiple users without severe degradation in the performance of the system.
FDMA and TDMA are the well-known multiplexing techniques used in wireless communication
systems [3].
OFDM is an advanced form of Frequency Division Multiplexing (FDM) where the frequencies
multiplexed are orthogonal to each other and their spectra overlap with the neighboring carriers [4]. In
a standard FDM system, the subcarriers do not overlap. OFDM is built on the principle of overlapping
orthogonal subcarriers. The frequency-domain view of the signal is shown in Figure 1. The peak of
one subcarrier coincides with the nulls of the other subcarriers due to the orthogonality. Thus, there is
no interference from other subcarriers at the peak of a given subcarrier even though the subcarriers
spectrums overlap. It can be understood that OFDM systems avoid the loss in bandwidth efficiency
prevalent in systems using non-orthogonal carrier sets. This brings in huge benefits in spectral
efficiency for OFDM systems over earlier systems [5].
Fig. 1. OFDM frequency spectrum [5]
In OFDM, the frequencies of the subcarriers are chosen so that the subcarriers are orthogonal to each
other, eliminating cross-talk between sub-channels, such that inter-carrier guard bands are not
required. Orthogonality is defined for both real and complex-valued functions. The functions  m (t)
and  n (t) are said to be orthogonal with respect to each other over the interval a < t < b if they satisfy
the condition:
b
 a
 m (t ) n (t ) dt  0
(1)
where n is not equal to m.
Two periodic signals are said to be orthogonal if they are mutually independent of each other and
when the integral of their product over one period is equal to zero. Signals S1 and S2 are considered
orthogonal if their inner product is zero.
T
( s1 s 2 )   s (t ) s
0
1 2 (t )dt  0 (2)
112
Orthogonal signals can be separated at the receiver by correlation techniques; hence, Inter Symbol
Interference ISI among orthogonal carriers can be eliminated. Orthogonality can be achieved by
carefully selecting carrier spacing, such as letting the carrier spacing be equal to the reciprocal of the
useful symbol period [6].
The orthogonality of the carriers means that each carrier has an integer number of cycles over a
symbol period. Due to this, the spectrum of each carrier has a null at the center frequency of each of
the other carriers in the system. This results in no interference between the carriers, although their
spectra overlap. The separation between carriers is theoretically minimal so there would be a very
compact spectral utilization.
OFDM systems have the advantage of high spectral efficiency and robustness to multipath
interference [7]. They are also attractive for the way they handle ISI, which is usually introduced by
frequency selective multipath fading in a wireless environment. Each subcarrier is modulated at a very
low symbol rate, making the symbols much longer than the channel impulse response. In this way, ISI
is diminished. Moreover, if a guard interval between consecutive OFDM symbols is inserted, the
effects of ISI can completely vanish. This guard interval must be longer than the multipath delay.
Although each subcarrier operates at a low data rate, a total high data rate can be achieved by using
a large number of subcarriers. ISI has little or no effect on OFDM systems. Hence, an equalizer is not
needed on the receiver side.
A well-known disadvantage of the OFDM system is its sensitivity to frequency offset between
transmitted and received signals, which may be caused by Doppler shift in the channel, or by the
difference between the transmitter and receiver local oscillator frequencies. The carrier frequency
offset causes loss of orthogonality, in which case the signals transmitted on each carrier are not
independent of each other, thus leading to Inter Carrier Interference (ICI) [8-9].
1.1 Typical OFDM System
Fig. 2 shows the block diagram of a typical OFDM transceiver [10]. The following sections discuss the
function of the main blocks.
Fig. 2. The basic block diagram of an OFDM system in AWGN channel
1.1.1 Signal Mapping
A large number of modulation schemes are available allowing the number of bits transmitted per
carrier per symbol to be varied. The main design issue of the modulator is the constellation used. The
constellation is the set of points that can be transmitted on a single symbol. The used constellation
affects several important properties of a communication system, Increasing the number of points in
the constellation does not change the bandwidth of the transmission, thus using a modulation scheme
with a large number of constellation points, allows for improved spectral efficiency. An OFDM system
113
performs modulation and demodulation for each subcarrier separately, although usually in a serial
manner, to reduce complexity [11].
1.1.2 Inverse Discrete Fourier Transform
In OFDM, the data at the transmitter is treated as a frequency-domain representation of the signal. It
is a sampled digital signal, and it is defined such that the discrete Fourier spectrum exists only at
discrete frequencies. To obtain a signal that can be transmitted, it has to be in the time domain. IDFT
can be easily applied to the transmitter and DFT can be applied at the receiver end to regain the
original data in frequency domain. Since the basis of the Fourier transform is orthogonal in nature, we
can implement it to get the time domain equivalent of the OFDM signal from its frequency
components. In practice, Fast Fourier Transformation for an N-input signal system is implemented,
rather than direct DFT and IDFT, because of their lower complexity and higher performance [8] [12].
1.1.3 Guard Interval and Cyclic Prefix
In a multipath environment, a transmitted symbol takes different times to reach the receiver through
different propagation paths. From the receiver's point of view, the channel introduces time dispersion
in which the duration of the received symbol is stretched. Extending the symbol duration causes the
current received symbol to overlap with previous received symbols and results in Inter Symbol
Interference (ISI). The effect of ISI on an OFDM signal can be improved by adding a guard period to
the start of each symbol. This guard period is a cyclic copy that extends the length of the symbol
waveform [13].
Fig. 3 shows the insertion of a guard period. The total length of the symbol is Ts=Tg+Tfft, where Ts is
the total length of the symbol in samples, Tg is the length of the guard period in samples, and Tfft is the
size of the IFFT used to generate the OFDM signal.
Fig. 3. Guard period insertion in OFDM
1.1.4 Communication Channel
The main sources of noise are thermal background noise, electrical noise in the receiver amplifiers,
and inter-cellular interference. Noise is also generated internally in the communication system as a
result of Inter Symbol Interference (ISI) and Inter Carrier Interference (ICI). These sources of noise
decrease the Signal to Noise Ratio (SNR), ultimately limiting the spectral efficiency of the system.
Noise, in all its forms, is the main detrimental effect in most radio communication systems.
Noise in radio communication systems is commonly modeled as AWGN. This noise has a uniform
spectral density (making it white), and a Gaussian distribution in amplitude, also referred to as a
114
normal distribution. Thermal and electrical noise from amplification, primarily has white Gaussian
noise properties, allowing them to be modeled accurately with AWGN [1].
1.2 Effect of Carrier Frequency Offset
In an AWGN channel, the frequency offset is modeled as a time-varying multiplicative factor

introduced in the channel as shown in Fig. 4.
w(n)
x(n) y(n)
Fig. 4. Frequency Offset Model
The received signal is given as:
(3)
where ε is the normalized frequency offset and w(n) is the AWGN introduced in the channel.
The effect of this frequency offset on the received symbol stream can be understood by considering
th
the received symbol on the k subcarrier Y(k)
N 1
Y ( k )  X ( k ) S ( 0)   S (l  k )X (l )  n
l  0,l # k
k for k  0,1,...., N  1 (4)
where N is the total number of subcarriers, X(k) is the transmitted symbol (M-ary phase-shift keying
th
(M-PSK), for example) for the k subcarrier, nk is the FFT of w(n) and S(l-k) terms are the complex
coefficients for the ICI components in the received signal. X(k)S(0) represents the desired signal. if
there is no frequency error, this term takes its correct value. The succeeding terms define the ICI
components in the received data.
th
The ICI components are the interfering signals caused by transmitted subcarriers other than the k
subcarrier. The complex coefficient S(l - k) is given by
(5)
To illustrate the effect of CFO synchronization error, the frequency spectrum of OFDM signal with 5
subcarriers is shown in Figure 5 for ϵ = 0 and 0.2. The dotted lines with the circle markers represent
the signal at DFT output. It is obvious from Figure 5(a) that the DFT output only consists of the signal
transmitted to the corresponding subcarrier without any interference from the neighboring subcarriers.
However, for ϵ = 0.2 as shown in Figure 5(b), the DFT output consists of both the signal transmitted
on the corresponding subcarrier and interference from the transmitted signal on the neighboring
subcarriers [9].
The main task of synchronization is to estimate and compensate for the frequency offset (ε).
115
Fig. 5. The effect of CFO on the subcarriers orthogonally 9]
1.3 Factors Inducing ICI
ICI is a special problem in the OFDM system. In this section, the factors inducing ICI will be
investigated. This frequency offset is caused by synchronization errors and/or Doppler shifts.
Synchronization errors arise from differences between the frequencies of the local oscillators at the
transmitter and receiver. Doppler shifts are caused by relative motion between the transmitter and
receiver.
In addition, multipath fading affects ICI. Actually, the multipath fading does not cause ICI, but it
aggravates the ICI problem. In a multipath fading medium, ICI is more complicated to calculate,
because there are many time-delayed versions of received signals with different gains and different
phase offsets.
In this chapter, our main concern is the normalized carrier frequency offset (CFO) which can be
expressed as
=∆ / =∆ (6)
where ε is the normalized CFO, fsub denotes the subchannel spacing, Ts is the subcarrier symbol
period and Δf is carrier frequency offset (CFO). Δf can be expressed by:
∆ = ync+ (7)
where fD is Doppler shift and fsync is the synchronization error between transmitter and receiver. These
effects are explained, in detail, in the following subsections.
1.3.1 Doppler Effect
Doppler effect occurs when there is a relative motion between the receiver and transmitter. In general,
the Doppler frequency shift is expressed in terms of the relative velocity, the angle between the
velocity direction and transmitter, and the carrier frequency [9].
The Doppler rate gives information about how fast the channel is varying compared to the data rate.
In general, the Doppler frequency, fD is given by:
f D  f D max cos (8)

vfc
f D m ax  (9)
c
116
Where  is the angle between relative velocity vector and that of the signal direction; generally
modeled as a uniformly distributed random variable between 0 and 2π, fDmax is the maximum Doppler
frequency, v is the velocity of the receiver relative to transmitter, fc is the carrier frequency and c is the
propagation velocity of the transmitted signal (speed of light). In mobile radio channels, the maximum
Doppler spread is an important parameter used in adaptation, to optimize the system’s performance,
enhance its capacity and utilize available resources efficiently.
1.3.2 Synchronization Errors
In general, most wireless receivers cannot achieve perfect frequency synchronization. Practical
oscillators at the receiver and transmitter operate at the same nominal frequency, but are never
identical. This effect leads to frequency offset. Although this small offset may be negligible in some
communication systems, it is a severe problem in OFDM systems.
The IEEE 802.11a standard requires the oscillators to have frequency errors within 20 ppm (or 20 X
-6
10 ). For a carrier of 5 GHz, this means a maximum frequency error of 200 KHz.
This error is relatively large compared to the frequency spacing of the carrier, which is only 312.5 KHz
[IEEE 802.11a with Channel bandwidth of 20 MHz and 64 subcarriers]. Hence, the frequency offset
cannot be ignored [7].
1.4 Methods of ICI Reduction
The offsets between the frequency reference of the transmitter and receiver give rise to ICI which in
turn degrades the performance of the OFDM system. To reduce this ICI, several researchers
developed different techniques to mitigate the frequency offsets as follows [14]:
1.4.1 Time Domain Windowing
If the signal is transmitted in a band-limited channel, certain portions of the signal spectrum will be cut
off, leading to inter carrier interference. To diminish this interference, the spectrum of the signal
waveform needs to be more concentrated. This is achieved by windowing the signal. Basically,
windowing is the process of multiplying the transmitted signal waveform by a suitable function. The
same window is used at the receiver side to retrieve the original signal. The ICI will be eliminated if
the product of the window functions satisfies the Nyquist symmetry criterion. Nyquist windows reduce
the side lobes, which in turn provide the carrier orthogonality. However, windowing is done symbol by
symbol, consequently reducing the spectral efficiency to a large extent [15].
1.4.2 Frequency Domain Equalization
The fading distortion in the channel causes ICI in the OFDM demodulator. The pattern of ICI varies
from frame to frame for the demodulated data but remains invariant for all symbols within a
demodulated data frame. Compensation for fading distortion in the time domain introduces the
problem of noise enhancement. So, the frequency domain equalization process is approached for the
reduction of ICI by using suitable equalization techniques. We can estimate the ICI for each frame by
inserting frequency-domain pilot symbols in each frame, as shown in Figure 6 [16].
Fig. 6. Pilot subcarrier arrangement.
117
This technique can only reduce the ICI caused by fading distortion which is not the major source of
ICI. The major source of ICI is the frequency mismatch between the transmitter and receiver. The
above method cannot address it. Again, it is only suitable for flat fading channels, but in mobile
communication, the channels are frequency selective fading in nature because of multipath
components. Here, also, the channel needs to be estimated for every frame. Channel Estimation is
complex, expensive and time consuming.
1.4.3 Pulse Shaping
Each carrier in the OFDM spectrum consists of the main lobe followed by many side lobes with
decreasing amplitudes. As long as orthogonality is maintained, there is no interference among the
carriers because, at the peak of every carrier, all other carriers have a spectral null. Hence, the
individual carrier is easily separated. When there is a frequency offset, the orthogonality is lost
because now the spectral null does not coincide with the peak of the individual carriers. The ICI power
will go on increasing as the frequency offset increases. The purpose of pulse shaping is to reduce the
side lobes. If we can reduce the side lobe significantly, then the ICI power will also be reduced
significantly. The significant pulse shaping functions are the raised cosine pulse (RC), the better than
raised cosine pulse (BTRC), the SINC power pulse (SP) and the improved SINC power pulse (ISP)
[17].
1.4.4 Repeated Data Methods- ICI self-cancellation schemes
Several studies advocated repeated data transmission as a solution for ICI problems. Among those
ICI suppression schemes, much attention has been paid to the so-called ICI self-cancellation
schemes due to their simplicity. The principle of this technique is to map each transmitted symbol onto
a group of r (r≥2) subcarriers, as shown in Figure 7 and the ICI generated within the group can be
self-canceled through the characteristics of ICI coefficients.
Fig. 7. Basic block diagram of an ICI self-cancellation OFDM systems
The first proposed technique was reported by Zhao [18], and is known as Adjacent Symbol Repetition
(ASR). It depends on repeating the same symbol with opposite polarity on two adjacent subcarriers, k
and (k+1), and then combining the received samples at the receiver.
The Phase-Rotated Data Allocation (PRDA) scheme [19] is closely related to ASR, where the same
symbol is repeated with a phase shift - π/2 on two adjacent subcarriers, k and (k+1). An improved
form of PRDA was proposed in [20] where each symbol is repeated with different complex weights on
four consecutive subcarriers, which impacts the throughput severely.
In the Symmetric Symbol Repetition (SSR) scheme [21], the repeated symbols with opposite polarity
are transmitted on subcarrier k and (N - k - 1). Therefore, the data block becomes x = ( X0, X1,...,
118
XN/2+1 , - XN/2+1,….., -X1 , -X0 ) .The receiver combines the received samples Yk and YN-k-1 and the
decision variable is obtained.
While ASR and SSR schemes are effective in reducing the ICI term on the received signal, the
received signal still experiences the same phase rotation as a conventional OFDM.
th
Another technique, known as Adjacent Conjugate Symbol Repetition (ACSR) [22], the symbol X on k
th
subcarrier is conjugated and repeated on (k +1) subcarrier. Therefore, the data block will be x = (X0 ,
X*0, ...XN/2+1 , X*N/2+1),where * denotes complex conjugation. The receiver combines the received
samples Yk and Yk+1
th
In Symmetric Conjugate Symbol Repetition (SCSR) [22], the modulated symbol X is mapped on k
th
and (N-1-k) subcarriers. Therefore, the data block becomes x = ( X0 , X1 ,…, XN/2+1 ,……., X*N/2+1 ,
' '
X*1 , X*0 ) . The receiver combines the received samples Y k and Y N -1- k .
*
Another form of repetition; (Xk,Xk+1=-X k), is also considered in the Peak-to-average power ratio
' '
(PAPR) scheme [23], where the receiver combines the received samples Y k and Y k+1. The PAPR
technique is closely related to the ACSR technique.
Unlike other ICI self-cancellation algorithms employing data repetition or conjugate data repetition
within the same OFDM symbol interval, the Conventional Conjugate Cancellation (CC) Scheme
transmits two independent paths such that the ICI weighting coefficient for one path could have an
opposite polarity of that for the other path. The CC scheme achieves high CIR values when the
frequency offset is small. However, the CIR curve of the CC scheme declines more rapidly than those
of other ICI cancellation schemes. This property makes the CC scheme undesirable if the frequency
offset is high.
The purpose of the General Phase Rotated Conjugate Transmission (PRCC) scheme is to enhance
the performance of the CC scheme at both low and high frequency offset conditions [24]. Like the CC
scheme, this scheme adopts two-path transmission, but introduces a new variable, φ. It allows the
transmitter to tune the transmitted signals so that the ICI effects of the two-paths 'signals could be
mutually canceled at the receiver. For the general PRCC scheme, the first path employs the standard
OFDM signal with artificial phase rotation of φ, while the second path adopts the conjugate of the
jφ −jφ *
standard OFDM signal with artificial phase rotation of −φ, i.e., e Xn and (e Xn) , respectively.
Some popular ICI cancellation schemes including ICI self-cancellation, ICI conjugate cancellation and
general phase rotated conjugate cancellation are summarized in Table 1.1.
Other relevant work utilizing ICI self-cancellation schemes was proposed and implemented for
wavelet-based OFDM systems [cf. for example Ref. 25, 26]. Alternative ICI self-cancellation schemes
based on differential encoding and decoding were also depicted in the literature [cf. for example
Ref.27].
119
Table 1. 1 Different ICI Self- Cancellation Schemes
Schemes Description CIR Phase Rotation

Adjacent TX Signal 2 S (0)  S (1)  S (1)
2
CIRASR  Im[ 2S (0)  S (1)  S (1)]

Symbol x = ( X0, -X0, X1, - N 1   tan 1
Repetition X1,……………….., XN/2+1, -
 2S (l )  S (l  1)  S (l  1)
2
Re[ 2 S (0)  S (1)  S (1)]
(ASR) XN/2+1 ). i  2 , 4 ,..
RX Signal
Y ' (k )  Y ( K )  Y ( K  1)
Symmetric TX Signal 2S (0)  S ( N  1)  S (1  N )
2
Im[ 2S (0)  S ( N  1)  S (1  N )]
CIRSSR    tan 1
Symbol x = ( X0, X1,..., XN/2+1, - N / 2 1 Re[ 2S (0)  S ( N  1)  S (1  N )]
 S (l )  S (l )  S ( N  l  1)  S (l  1  N )
2
Repetition XN/2+1,….., -X1, -X0 )
i 1,..
(SSR) RX Signal
Y ' (k )  Y (k )  Y ( N  k  1)
Adjacent TX Signal Im[S (0)  S * (0)]
2
S (0)  S * (0)
Conjugate x = ( X0, X*0, X1, X*1,…… CIRACSR    tan 1
N 1 N 1
Re[ S (0)  S * (0)]
  S (l  1)  S
2 2
Symbol ……….., XN/2+1, X*N/2+1 ) S (l )  S * (l )  *
(l  1)
i  2 , 4 ,.. l 0, 2
Repetition RX Signal
(ACSR) Y ' (k )  Y (k )  Y (k  1)
*
Symmetric TX Signal Im[S (0)  S * (0)]

2
S (0)  S * (0)
Conjugate x = ( X0, X1,…, XN/2+1, …., CIRSCSR    tan 1
Symbol X*N/2+1, X*1, X*0 )

N / 2 1

N 1
S (l )  S * (l )   S ( N  l  1)  S * (l  1  N )
2 2
Re[ S (0)  S * (0)]
i 1,.. l 0, 2
Repetition RX Signal
(SCSR) Y ' (k )  Y (k )  Y * ( N  1  k )
PAPR TX Signal S (0)  S (1)  S (1)  S (0)
2 2
Im[ S (0)  S (1)  [ S (1)  S (0)]]
Reduction Xk, Xk+1=-X*k CIRPAPR  N 2
  tan 1
Re[ S (0)  S (1)  [ S (1)  S (0)]]
 S (l )  S (l  1)
2
 S (l  1)  S (l )
2
Scheme in RX Signal
OFDM Y '' (k )  Y ( K )  Y * ( K  1) l  even
Systems
120
Schemes Description CIR Phase Rotation

ICI Conjugate TX Signal
S ( 0 )  S * ( 0)
2 Im[ S (0)  S * (0)]
Cancellation Two conjugated frames   tan 1
Method(CC) RX Signal CIR cc  N 1 Re[ S (0)  S * (0)]
 S (l )  S * 2
Y "k  (Yk  Y 'k ) / 2 (l )
i 1,..
General TX Signal Im[e  j S (0)  e j S * (0)]

2
e  j S (0)  e j S * (0)   tan 1
Phase ejφXn and (e−jφXn)* CIR PRCC 
Rotated RX Signal N 1 Re[e  j S (0)  e j S * (0)]
e  j 2
S (l )  e j S * (l )
Conjugate Y "k  (Yk  Y 'k ) / 2 n 1,..
Transmission
(PRCC)
Phase- TX Signal Im[S (0)  S (1)  e  j / 2 [S (1)  S (0)]]
2
S (0)  S (1)  e  j / 2 [ S (1)  S (0)]
Rotated Data X(k),X(k+1)= e-j/2 X(k) CIR PR    tan 1
N 2
Re[S (0)  S (1)  e  j / 2 [S (1)  S (0)]]
 S (l )  S (l  1)  e  j / 2 2
Allocation RX Signal [ S (l  1)  S (l )
i 0
(PRDA) Y ' (k )  Y ( K )  e  j / 2Y ( K  1) even
 CIR (Carrier to Interference) is the ratio of the signal power to the power in the interference components. CIR serves as a good indication of signal quality. While
deriving the CIR expression, the additive noise is omitted, and it is assumed that the transmitted data have zero mean and are statistically independent.
 For an OFDM system, the phase rotation of the desired signal X(k) caused by the frequency offset is
 Im(S0 ) 
  tan 1  
 Re( S0 ) 
where S0 is given by equation (5) when l=k.
121
2. PROPOSED ICI SELF -CANCELLATION SCHEMES
New ICI self-cancellation schemes are proposed to mitigate the effect of ICI on OFDM systems. The
main idea of these schemes is first introduced as a simple single carrier system. Then it is applied to
multicarrier modulation systems.
2.1 Single-Carrier Transmission
2.1.1 Conjugate Multiplicative Single-Carrier Transmission (CM-SC)
In CM-SC, each symbol is conjugated and repeated to build a data pattern in the transmitted stream.
This data pattern will be utilized at the receiver to cancel the CFO errors. Thus, the transmitted stream
* * *
takes the form X1X1 X2X2 X3X3 …..
At the receiver, each symbol suffers a CFO phase error φ(t) = 2ft. Consequently, the received
*
symbols Y1 and Y2 corresponding to transmitted symbols X1 and X1 are given by (in absence of
noise):
j(θ1 + φ ) -j(θ1 -φ )
Y1 = r1 e & Y2 = r1 e (10)
where r1 and θ1 are the magnitude and the phase of the received symbol corresponding to transmitted
signal X1. In order to eliminate the effect of CFO phase shift completely, the output of the receiver
may be calculated as follows:
 Calculate Y2*.
 Multiply Y1 by Y2*. Note that the offset phase angle φ is eliminated in the product.
θ
Calculate the received signal as (11)
Eqn. (11) indicates that the CFO error is completely eliminated.
Our proposed ICI self-cancellation scheme, therefore, is labeled the conjugate multiplicative scheme,
as opposed to all previously proposed ICI self-cancellation schemes which are additive in nature.
Although the CM-SC technique eliminates the CFO error completely, the multiplication process has
several disadvantages. First, constellation points of the lower half of the complex plane are folded
over those of the upper half, leaving only half the constellation points usable. Also, multiplication
complicates the symbol detection process considerably. It also adds significant hardware complexity.
Additionally, the multiplication operation impacts the SNR ratio negatively [28, 29].
2.1.2 Conjugate Phase Additive Single-Carrier Transmission (CPA-SC)
We modified the above algorithm as follows:

*
 Calculate phases of Y1 and Y2 .
 Find the average of these two phases θav.
θ
 Calculate the received signal as
The proposed ICI self-cancellation scheme (CPA-SC) is, therefore, labeled the Conjugate Phase
Additive (CPA) scheme. Note also that complete CFO elimination is still valid for all
modulation schemes, provided that AWGN effect is negligible and the system is a single carrier
system. Partial cancellation of CFO errors takes place if any of these two conditions is violated. Note
also that all the constellation points of any modulation scheme can be used without ambiguity. Also,
received signals are linearly related to transmitted signals, so simple decoding is recovered.
Additionally, averaging the amplitudes and phases of the received signals have a positive impact on
the SNR ratio.
122
2.2 MultiCarrier Transmission
In this section; the proposed modulation scheme introduced in section (2.1) for single-carrier
modulation is applied to OFDM systems. We name it the Conjugate Phase Additive MultiCarrier
Transmission.
2.2.1 Adjacent Conjugate Phase Additive OFDM (ACPA)
In ACPA-OFDM the modulated data symbol X is mapped on two subcarriers as X and X*, the
receiver combines the received samples Yk and Yk+1 by calculating the average phases of the two
adjacent subcarriers after conjugating Yk+1 as illustrated in Section 2.1.2 and Figure 8 [28, 29].
2.2.2 Symmetric Conjugate Phase Additive OFDM (SCPA)
In SCPA-OFDM the modulation symbol X is mapped as X and X* on the two subcarriers k and (N-k-
1), respectively. The receiver combines the received samples Yk and YN-k-1 by adding the average
phases of the two symmetric subcarriers after conjugating YN-k-1 as shown in Figure 9 [28, 29].
IFFT
Input Binary Modulator

Conjugate
Data
AWGN
0.5
Odd
FFT
 -
Even
0.5
RX Symbols Demodulator
0.5
Odd
 Even
0.5
Fig. 8. Proposed ACPA OFDM System model.
123
Symmetric IFFT
Input Data
Modulator
Conjugate
Binary Data
AWG
N
0.5
Odd
Anti- FFT
Symmetric  -
Data Even
0.5
Demodulator
RX Symbols
0.5
Odd
 +
Even
0.5
Fig. 9. Proposed SCPA OFDM System model
3. SIMULATION RESULTS AND PERFORMANCE EVALUATION
Computer simulation is an effective and efficient way to represent the actual performance of mobile
radio systems. We used MATLAB to simulate the proposed as well as conventional OFDM
communication systems
In this section, CPA-SC single carrier 16-QAM and 256-QAM schemes are implemented. The
proposed schemes show:
 Graceful degradation for high ε values.

 Very good performance at high modulation schemes (e.g. 256-QAM).
The reader can refer to [28, 29] for a discussion on the proposed scheme performance for low order
constellation schemes (BPSK, QPSK …etc.).
The proposed schemes are applied to multicarrier modulation OFDM systems. ACPA and SCPA are
implemented and simulated using MATLAB. They are compared with well-known ICI self–cancellation
techniques [18-24]. For higher modulation schemes and higher frequency offsets, the proposed
schemes show significant improvement in BER performance over standard OFDM and all the existing
ICI self-cancellation schemes.
3.1 Conjugate Phase Additive Single Carrier Transmission (CPA-SC)
3.1.1 CPA-SC 16-QAM System
A sequence of uniformly distributed pseudorandom numbers is generated using the rand MATLAB
built-in function, and hence, fed to the simulator as the incoming binary sequence. Bits are fed into a
16-QAM modulator to generate the modulated symbols. ICI Cancellation mapper is used to build a
* * *
data pattern in the transmitted stream in the form X1X1 X2X2 X3X3 …..
124
On the receiver side, the effect of ICI is eliminated by calculating the angle and magnitude of both odd
and even received symbols. Then, a 16-QAM demodulator is used to restore the generated input
data. The input data fed to the transmitter is compared with the output data from the receiver -after
adding AWGN channel effect- by dividing the total number of unequal pairs of data elements by the
total number of input data elements from one source.
Figure 10 shows BER vs. Eb/No for both CPA-SC 16-QAM with ICI self -cancellation and normal 8PSK
for different values of carrier frequency offset . As the frequency offset increases to more than 0.1,
BER of 8PSK degrades dramatically relative to CPA-SC QAM16. Actually, 8PSK lost the
synchronization. CPA-SC 16-QAM is better than 8PSK scheme in terms of AWGN noise and
frequency offset elimination. CPA-SC 16-QAM BER sensitivity to Eb/No is approximately 1 dB higher
than 8PSK at ε=0. This indicates that the BER of the CPA-SC scheme degrades more gracefully as
the frequency offset ε increases.
Bit error probability curve for CPA-SC 16-QAM & 8PSK OFDM
-1
10
Bit Error Rate
-2
10
CPA-SC 16-QAM  = 0
-3
10 CPA-SC 16-QAM  = 0.15
CPA-SC 16-QAM  = 0.25
8PSK  = 0
8PSK  = 0.15
-4
8PSK  = 0.25
10
2 4 6 8 10 12 14
Eb/No, dB
Fig. 10. BER versus Eb/No for CPA-SC 16-QAM and 8PSK system
3.1.2 CPA-SC 256-QAM System
Figure 11 shows BER vs. Eb/No for both CPA-SC 256-QAM with ICI self -cancellation and normal
128-QAM for different values of carrier frequency offset . As the frequency offset increases to more
than 0.05, the BER of QAM128 degrades dramatically relative to CPA-SC 256-QAM. CPA-SC 256-
QAM is better than the 128-QAM scheme in terms of AWGN noise and frequency offset elimination.
CPA-SC 256-QAM BER sensitivity to Eb/No, is approximately 2 dB higher than 128-QAM at ε=0. This
indicates that the BER of the CPA-SC scheme degrades more gracefully with frequency offset ε
increases.
125
Bit error probability curve for CPA-SC 256-QAM & 128-QAM OFDM
-1
10
Bit Error Rate
-2
10
CPA-SC 256-QAM  = 0
CPA-SC 256-QAM  = 0.02
CPA-SC 256-QAM  = 0.1
128-QAM  = 0
128-QAM  = 0.02
-3 128-QAM  = 0.1
10
2 4 6 8 10 12 14 16 18 20
Eb/No, dB
Fig. 11. BER versus Eb/No for CPA-SC 256-QAM and 128-QAM system
3.2 Conjugate Phase Additive OFDM Transmission
At the beginning of the simulation, all required OFDM parameters and program variables are
initialized. The variables include IFFT size (an integer of a power of two), Number of carriers, Signal-
to-Noise Ratio in dB, and Frequency offset .
3.2.1 Performance Analysis of Conjugate Phase Additive Multi Carrier Transmission (ACPA /
SCPA)
The ACPA scheme is compared with SCPA scheme using the CIR and phase rotation performance
measures.
3.2.1.1 CIR Performance
Figure 12 shows the CIR performance of the proposed ACPA and SCPA schemes in decibels as a
function of the normalized frequency offset ε, where N = 1024. It is evident that the two proposed
schemes, ACPA and SCPA, are identical in CIR analysis and give about 40dB to 10dB CIR
improvement for small to medium frequency offsets in the range 0<ε≤0.2.
CIR versus  for OFDM systems

120
Standard OFDM
100 ACPA
SCPA
80
CIR (dB)
60
40
20
-20
0 0.1 0.2 0.3 0.4 0.5
Normalized Frequency Offset 
Fig. 12. The CIR performance of the ACPA, SCPA schemes
126
Fig. 13 Compares the CIR performance of the SCPA scheme with other schemes published in
references [18-24]. It is clear from this figure that SCPA gives better CIR performance than the
standard and SCSR OFDM schemes. Although the ASR and SSR schemes give better CIR
performance than the proposed SCPA, they do not succeed in eliminating the phase rotation due to
CFO as is evidenced by Figures 14, 15].
CIR versus  for OFDM systems

90
Standard OFDM system
80
ICI SCPA
70 ICI SCSR
60 ICI SSR
ICI ASR
50
CIR (dB)
40
30
20
10
-10
-20
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Normalized Frequency Offset 
Fig. 13. CIR improvement using different ICI self-cancellation schemes
3.2.1.2 Phase Shift
In terms of phase rotation, Figure 14 shows that the normal OFDM and SSR model have a phase shift
o
of about 45 if the normalized frequency offset is 0.25. Instead, the proposed algorithms ACPA/ SCPA
show a much better performance and high immunity to frequency offset variation.
Fig. 14. Phase shift vs. normalized frequency offset for SSR, ACPA -SCPA schemes
127
3.3 BER Performance Comparison with Other Schemes
Several simulations were done to verify the effectiveness of the proposed ICI self-cancellation
schemes. The performance of the proposed ICI self-cancellation scheme is evaluated using
QPSK,16-QAM and 256-QAM.
3.3.1 QPSK system
As illustrated in Figure 15, for higher frequency offsets, the SCPA-QPSK scheme gives a better SNR
gain than SCSR by 2dB as noticed at BER 10 and  = 0.2. Also, for lower  values, the SCPA-QPSK
-3
scheme behaves better than ASR by SNR gain of 1.5dB as shown at BER 10 and  = 0.1. Although
-3
ASR offers good performance at higher ε, the next sections demonstrate that this ASR performance
falls substantially below that of SCPA for higher modulation schemes.
Bit error probability curve for QPSK SCPA & SCSR & ASR OFDM
-1
10
-2
Bit Error Rate
10
-3
10
SCPA  =0.1
SCPA  =0.2
SCSR  =0.1
SCSR  =0.2
ASR  =0.1
ASR  =0.2
-4
10
2 4 6 8 10 12 14
Eb/No, dB
Fig. 15. BER comparison of SCPA, SCSR and ASR QPSK Scheme for ε = 0.1 & 0.2
3.3.2 16-QAM system
The BER performance of standard OFDM, ASR, SCSR, and the proposed ICI self-cancellation
scheme SCPA in AWGN channel for  = 0.05 and 0.15 are shown in Figure 16. The modulation
scheme is 16-QAM, and the FFT symbol size =64. This figure reveals that for 16-QAM, the
performance of all the systems degrades if the CFO increases. But among all the schemes the
proposed scheme offers much improved performance. For example, SCPA offers an SNR gain of 5
dB over that of SCSR at BER of 10 and  = 0.15. SCPA gives higher immunity against CFO errors at
-2
ε = 0.15 compared to the SCSR scheme, while the ASR scheme performance fails completely at ε =
0.15.
128
Bit error probability curve for 16-QAM SCPA & SCSR & ASR OFDM
-1
10
Bit Error Rate
-2
10
SCPA  = 0.05
SCPA  = 0.15
SCSR  = 0.05
SCSR  = 0.15
ASR  = 0.05
-3
ASR  = 0.15
10
2 4 6 8 10 12 14
Eb/No, dB
Fig. 16. BER comparison of SCPA, SCSR and ASR 16-QAM Scheme for ε = 0.05 & 0.15
3.3.3 256-QAM System
The BER performance of ASR, SCSR and the proposed ICI self-cancellation SCPA scheme in an
AWGN channel for = 0.05 and 0.1 is shown in Figure 17. The modulation scheme is 256-QAM, and
FFT symbol size = 256. The proposed scheme offers an SNR gain of around 7 dB over that of SCSR
at BER of 10 and  = 0.1.
-1
It is noticeable that the SCPA gives good performance and higher immunity than the SCSR scheme
against CFO errors for ε = 0.1, while ASR scheme fails completely even with small ε
Bit error probability curve for 256-QAM SCPA & SCSR & ASR OFDM
-1
10
Bit Error Rate
-2
10
SCPA  =0.05
SCPA  =0.1
SCSR  =0.05
SCSR  =0.1
ASR  =0.05
ASR  =0.1
-3
10
5 10 15 20 25
Eb/No, dB
Fig. 17. BER comparison of SCPA, SCSR and ASR 256-QAM Scheme for ε = 0.05 & 0.1
129
4. HARDWARE IMPLEMENTATION
One of the principle challenges of OFDM systems is the cost and complexity of the system
implementation. Field Programmable Gate Arrays (FPGA) provide a viable platform for this purpose
because of their low-cost, easy programmability and vast resources [30].
The Xilinx Spartan 3A DSP FPGA Video Starter Kit (VSK) is a development platform consisting of the
Spartan-3A DSP 3400A Development Platform, the FMC-Video daughter card, and a VGA camera. It
is comprised of a variety of hardware and software components. The key features of the Spartan-3A
DSP 3400A starter Kit are:
 Spartan-3A DSP XC3SD3400A device has ample resources (for example, 126 embedded
DSP blocks) for implementing high-performance processing systems.
 2-line, 16-character LCD screen, PS/2 mouse and PS/2 keyboard port
 DB9 (RS232) connector/ DVI connector
 LEDs, DIP switches
4.1 Design Methodology
System Generator, developed by Xilinx to facilitate the design and implementation of DSP functions
on its FPGAs [31], is used as a graphical environment for creating hardware models for wireless
communication systems.
4.1.1 System Generator
Simulink provides a powerful high-level modeling environment for DSP systems and consequently is
widely used for algorithm development and verification. System Generator maintains an abstraction
level very much in keeping with the traditional Simulink block sets, but at the same time automatically
translates designs into hardware implementations that are faithful, synthesizable, and efficient.
The implementation using the Xilinx system generator is made efficient through the instantiation of
intellectual property (IP) blocks that provide a range of functionality from arithmetic operations to
complex DSP functions. These IP blocks have been carefully designed to run at high speed and to be
area efficient. In System Generator, the capabilities of IP blocks have been extended automatically to
fit gracefully into a system-level framework. For example, although the underlying IP blocks operate
on unsigned integers, System Generator allows signed and unsigned fixed-point numbers to be used,
including saturation arithmetic and rounding. User-defined IP blocks can be incorporated into a
System Generator model as black boxes which will be embedded by the tool into the HDL
implementation of the design [32].
4.1.1.1System Generator Design Flow
Simulink provides a graphical environment for creating and modeling dynamic systems. System
Generator consists of a Simulink library called the Xilinx Blockset, and software to translate a Simulink
model into a hardware realization of the model. System Generator maps system parameters defined
in Simulink (e.g. as mask variables in Xilinx Blockset blocks), into entities and architectures, ports,
signals, and attributes in a hardware realization. In addition, System Generator automatically
produces command files for FPGA synthesis, HDL simulation, and implementation tools, so that the
user can work entirely in graphical environments as he/she proceeds from system specification to
hardware realization [31]. The System Generator design flow is shown in Figure 18.
130
Fig. 18. System Generator design flow diagram
System Generator translates the Simulink model into a hardware realization by mapping Xilinx
Blockset elements into IP library modules, inferring control signals and circuitry from system
parameters (e.g. sample periods), and converting the Simulink hierarchy into a hierarchical VHDL
netlist.
4.1.2 Hardware/Software Co-simulation
System Generator provides hardware/ software co-simulation, making it possible to incorporate a

design running in an FPGA directly into a Simulink simulation [33,34]. "Hardware/ Software Co-
Simulation" compilation targets automatically create a bitstream and associate it with a block as
shown in Figure 19. When the system design is simulated in Simulink, results for the compiled portion
are calculated in actual FPGA hardware, verifying the functional correctness of the hardware. System
Generator for DSP supports Ethernet as well as JTAG communication between a hardware platform
and Simulink.
Fig. 19. FPGA based Hardware-Software (HW-SW) co-simulation environment
131
4.2 Hardware Implementation schemes
4.2.1 Proposed 256-QAM ACPA OFDM System
The proposed 256-QAM ACPA OFDM system, shown in Figure 20, is implemented using Simulink
and Xilinx blockset library. The block diagram of the system consists of four main blocks
- 256-QAM Modulator and Demodulator

- ICI Cancellation Mapping and De-mapping
- IFFT and FFT Processors.
- Channel block.
Fig. 20. 256-QAM ACPA OFDM scheme Block diagram
The 256-QAM ACPA ICI Cancellation mapping and De-mapping are implemented using FPGA as
shown in Figure 21 & Figure 22 respectively.
Fig. 21. 256-QAM ACPA ICI Cancellation mapping system generator blocks
132
Fig. 22. 256-QAM ACPA ICI Cancellation de-mapping system generator blocks
4.2.2 128-QAM with CFO Estimation and Correction Technique
The 128-QAM CFO-EC OFDM system, shown in Figure 23, is implemented using Simulink and Xilinx
blockset library. The block diagram consists of four main blocks
- 128-QAM Modulator and Demodulator

- CFO Estimation and Correction block
- IFFT and FFT Processors.
- Channel block.
The CFO Estimation and correction system generator blocks are implemented using FPGA as shown
in Figure 24.
Input Binary
QAM128
IFFT Add GI Channel
Data
Modulator
CFO
Estimation
Output QAM128 Remove CFO

FFT
Demodulator GI
Data Correction
Fig. 23. 128-QAM CFO-EC OFDM scheme Block diagram
133
Fig. 24. 128-QAM CFO Estimation & correction system generator blocks
4.3 Hardware Results and Performance Evaluation
4.3.1 Test Bench Setup
A test bench was developed using the Spartan-3A DSP 3400A FPGA Development Video Starter Kit.
System Generator provides a generic interface that uses JTAG and a Xilinx programming cable (e.g.
Platform Cable USB) to communicate with FPGA hardware. Figure 25 shows the hardware setup
model with the JTAG-based hardware co-simulation block.
Fig. 25. Setup for FPGA co-simulation test bench
134
4.3.2 Hardware Resources
A comparison was done between 256-QAM ACPA ICI Self–Cancellation scheme and the classical
128-QAM CFO–Estimation and Correction scheme (CFO-EC). The implementation results on
Spartan3A DSP are shown in Table 2. This table compares the hardware resources required for CFO
error correction in the classical CFO-EC scheme and the proposed ACPA scheme. The proposed
ACPA scheme requires less than 50% of the hardware resources needed to build the CFO error
correction function in the classical system.
Table 2. Comparison between the Hardware Resource Used to Correct the CFO Errors for the
proposed 256-QAM ACPA and the reference 128-QAM CFO -EC schemes
Logic Utilization OFDM 128-QAM with CFO-EC OFDM 256-QAM ACPA

No of Slice Flip Flops 5,426 2,193
No of 4 input LUTs 7,162 2,633
No of occupied Slices 3,985 1,583
4.3.3 Frequency Offset Analysis
The reader can refer to [28, 29] for a discussion on the proposed scheme hardware/ software co-
simulation for low order constellation schemes (BPSK, QPSK …etc.). In this section we consider only
the performance for the high order QAM256 modulation scheme.
For hardware co-simulation testing purposes, the data was modulated using 256-QAM in the ACPA
model and using 128-QAM in the CFO-EC system. Additive White Gaussian noise (AWGN) was
added, corresponding to Eb/No of 50dB. A frequency offset was applied to the data in the time domain.
Figure 26 shows BER curves versus the normalized frequency offset ε for the OFDM 128-QAM CFO-
EC system and the proposed 256-QAM ACPA system. A reference QAM128 OFDM system without
any CFO handling scheme is also included. The ACPA compares well to the 128-QAM CFO-EC
system.
Fig. 26. BER vs. ε for 128-QAM - EC and 256-QAM ACPA OFDM
135
5. CONCLUSION
In this chapter, a new ICI self-cancellation scheme is introduced to reduce the effect of ICI on OFDM
systems. The main idea of this scheme is first introduced using a simple single carrier system, and
then it is applied to multicarrier modulation systems. In the proposed scheme, each symbol is
conjugated and repeated to build a data pattern in the transmitted stream. This data pattern will be
utilized at the receiver to cancel the CFO errors. Thus, the transmitted stream takes the form
* * *
X1X1 X2X2 X3X3 …. At the receiver, each symbol suffers a CFO phase error. By calculating the
* *
phases of the received symbols Y1 and Y2 corresponding to transmitted symbols X 1 and X1 ,
averaging these two phases, the corrected received signal is calculated. Our proposed ICI self-
cancellation scheme is, therefore, labeled as Conjugate Phase Additive (CPA) scheme. We apply the
proposed CPA scheme to single carrier modulation systems and to OFDM systems. We propose two
variants of this system; the Adjacent Conjugate Phase Additive OFDM (ACPA) and the Symmetric
Conjugate Phase Additive OFDM (SCPA).
The proposed CPA ICI self-cancellation schemes offer low HW complexity. They display graceful
degradation BER rates as the frequency offset increases, especially for high modulation schemes.
Compared to the best published ICI self-cancellation scheme (ASR, SCSR), the ACPA scheme yields
7 dB improvement over that of SCSR at BER of 10 and  = 0.1 for 256-QAM modulation type.
-1
Finally, 256-QAM ACPA OFDM ICI self-cancellation system is implemented using Xilinx System
Generator tool for DSP design, co-simulated on Spartan3A DSP FPGA. The widely used CFO error
estimation and correction technique using cyclic prefix is also implemented on the same platform. The
two system are then compared in terms of hardware resource usage and frequency offset analysis.
The hardware required for the ACPA scheme is less than 50% of that required for the classical CFO-
EC scheme. Relative to the classical 128-QAM OFDM system with CFO-EC correction, the 256-QAM
ACAP scheme achieves comparable immunity against frequency offset errors over the practical
frequency offset.
COMPETING INTERESTS
Authors have declared that no competing interests exist.
REFERENCES
1. Haykin S. Communication Systems, 4th ed., Wiley; 2001.

2. Proakis J. Digital Communications, 5th ed., McGraw-Hill; 2007.
3. Rappaport TS. “Wireless Communication: Principles and Practice,” Prentice Hall, 2nd Edition;
2002.
4. Chang RW. “Synthesis of Band Limited Orthogonal Signals for Multichannel Data
Transmission,” Bell Syst. Tech. J. 1996; 45:1775-1796.
5. Heiskala J, Terry J. “OFDM Wireless LANs: A Theoretical and Practical Guide,” Sams,
December; 2001.
6. Litwin L, Pugel M. “The Principles of OFDM,” RF Design; 2001.
7. Intini AL. “Orthogonal Frequency Division Multiplexing for Wireless Networks,” University of
California, Santa Barbara, December; 2000.
8. Henkel W, Odling G, Borjesson P. “The Cyclic Prefix of OFDM/DMT – An Analysis,”
International Zurich seminar on broadband communications, ETH Zurich, 21February; 2002.
9. Van Nee R, Prasad R. “OFDM for Wireless Multimedia Communications,” Artech House,
January; 2000.
10. Taub H, Schilling DL, Saha G. “Principles of Communication Systems,” Tata McGraw Hill; 2008.
11. Dash A. “OFDM Systems and PAPR Reduction Techniques in OFDM Systems,” Bachelor
Thesis, Department of Electronics and Communication Engineering, National Institute of
Technology, Rourkela; 2006 – 2010.
136
12. Prasad R, Velez FJ. "WiMAX Networks Techno–Economic Vision and Challenges," Springer;
2010.
ISBN: 978-90-481-8751-5.
13. Dahlman E, Parkvall S, Sköld J. “3G Evolution-HSPA and LTE for Mobile Broadband,”
Academic Press, 2nd edition; 2008.
14. Kumar K, “Inter Carrier Interference Cancellation in OFDM Systems,” MSc. Thesis, National
Institute of Technology Rourkela; 2103.
15. Muschallik C, "Improving an OFDM reception using an adaptive Nyquist windowing," Consumer
Electronics, IEEE Transactions. 1996:259-269.
16. Ahn J, Lee HS. ―Frequency domain equalization of OFDM signal over frequency nonselective
Rayleigh fading channels, Electron. Lett. 1993; 29(16):1476–1477.
17. Moghaddam NM, Mohebbi M. “ICI Reduction Methods in OFDM Systems," Recent Advance in
Wireless Communication and Networks, Prof. Jia-Chin Lin (Ed.), ISBN: 978-953-307-274-6,
August 23; 2011.
18. Zhao Y, Haggman SG. “Intercarrier interference self-cancellation scheme for OFDM mobile
communication systems,” IEEE Trans. Commun. 2001; 49(7):1185–1191.
19. Yang CY, Lee GR, Wen JH. “ICI self-cancellation scheme with phase-rotated data allocation in
OFDM systems,” International Conference on Advanced Information Technologies (AIT), April,
Taichung, Taiwan; 2010.
20. Rahman M, Dey PK, Rabbi MF. "Improved ICI Self Cancellation Scheme for Phase Rotation
Error Reduction in OFDM System, "International Journal of Information and Electronics
Engineering. 2012; 2(2):581-584.
21. Sathananthan K, Rajatheva RMAP, Slimane SB. “Cancellation technique to reduce intercarrier
interference in OFDM,” IEE Elect.Lett. 2000;36:2078 -2079, Dec.
22. Sathananthan K, Athaudage C, Qiu B. “A Novel ICI Cancellation Scheme to Reduce both
Frequency Offset and IQ Imbalance Effects in OFDM,” Computers and Communications, 2004.
Proceedings. ISCC 2004. Ninth International Symposium. 2004; 2:708-713.
23. Fu Y, Kang SG, Ko CC. “A new scheme for PAPR reduction in OFDM systems with ICI Self-
Cancellation,” Vehicular Technology Conference. 2002:1418-1421.
24. Wang C, Huang YC. “Intercarrier Interference Cancellation Using General Phase Rotated
Conjugate Transmission for OFDM systems,” IEEE Trans. on Comm. 2010; 58(3):812-819.
25. Ayeswarya R, Prabha NA. Performance evaluation of ICI self-cancellation schemes in fractional
wavelet-based OFDM system. International Journal of Internet Technology and Secured
Transactions. 2020; 10(5):552-64.
26. Kaur N, Kumar N. Comparative analysis of ICI self cancellation techniques for wavelet OFDM
under different channels in simulink. Wireless Personal Communications. 2019; 105(4):1513-
25.
27. Ye J, Wang H, Yuan W, Wang N, Xu R. A Novel ICI Triple Cancellation Scheme in Fast Time-
varying Environment for OFDM Systems. In IOP Conference Series: Materials Science and
Engineering. 2018; 383(1):012055). IOP Publishing.
28. Reem I. Sayed, "New ICI Self Cancellation Scheme for OFDM Systems", Ph.D. thesis,
Electronics and Communications Department, Faculty of Engineering, Cairo University, Egypt.
Available from authors.
29. Habib SED, Reem I. Sayed, Hisham M. Hamed and Magdy Fikri, " New ICI Self Cancellation
Scheme for OFDM Systems", in International Journal of Computer and Communication
Engineering. 2014;3:1-5.
30. Hwang J, Milne B, Shirazi N, Stroomer J. “System Level Tools for DSP in FPGAs,” in Proc. XI
International Conference on Field Programmable Logic and Applications (FPL), Belfast,
Northern Ireland (UK); 2001.
31. Xilinx Web site. http://www.xilinx.com
32. Xilinx Core Generator System, http://www.xilinx.com/products/logicore/coregen/index. htm.
33. ISim HW Co-Simulation 14.1 Quick.
Available: http://www.xilinx.com/tools/feature/14_1_isim_hw_cosim_qrg.pdf
34. Areibi S. “Tutorial -Using Xilinx System Generator 13.4 for Co-Simulation on Digilent NEXYS3
(Spartan-6) Board”
137
Biography of author(s)
S. E. D. Habib
He received B.Sc. degree from Cairo University, Egypt in 1968, M. Sc. Degree in Solid State Physics from AUC, Egypt in 1971,
and M.Sc. and Ph.D. degrees in Electrical Engineering from University of Toronto, Canada, in 1973 and 1978 respectively. He
is currently an emeritus professor at the Electronics and Communication Department, Faculty of Engineering, Cairo University.
His current research interests include VLSI design, semiconductor devices. He has authored and co-authored three books, and
over 95 referred papers. He holds a British patent for his invention of the ALDMOST power MOST transistor. He supervised
and co-supervised 64 Ph.D. and M.Sc. theses. He was awarded the Egyptian State Encouragement Prize in 1985, the
Egyptian Legion of Distinction (first class) in 1995, and The Cairo University Appreciation Award for engineering sciences in
2016. He is a Life Senior Member of IEEE.
Dr. Reem I. Sayed

Electronics Factory, AOI, Egypt.
She received B.Sc. degree from Ain Shams University, Egypt in 2000. She received her M. Sc. and Ph.D degrees, both in
Electrical Engineering and communication from Cairo University, Cairo, Egypt in 2004 and 2014, respectively. Her current main
research interests include communications systems, VLSI, embedded systems design and IOT Based designs. She is the R&D
General Manager in Electronics Factory, AOI, Egypt, since January 2019.
Dr. Hisham M. Hamed

Wasiela Company, Egypt.
He received B.Sc with honors in 1989, MSc. in 1994, and PhD. in 2002, all in electronics and electronic communications, from
the Faculty of Engineering at Cairo University. He was the VLSI department manager at the R&D center, IEP, (Bahgat Group),
from April 15th 2001 to Feb 2003, where he supervised the design of an MPEG decoder chip. He was appointed assistant
professor of Electronics at the National Telecommunications Institute (NTI) in April 2003. From November 2005 to February,
2007 he was a consultant for SySDSoft, where he was involved in developing the baseband PHY, as well as the hardware
design, for a BlueTooth 2.0 chip. He is currently the chief technical officer of Varkon semiconductors, which he co-founded in
2008. His research interests include hardware implementations of DSP algorithms for communications applications, and
physical layer hardware design for wireless communications systems. He is a member of the Center for Wireless Studies, at the
Faculty of Engineering, Cairo University.
138
Prof. Magdi Fikri

He was born in Cairo, Egypt, 1942. He received his B. Sc. and M. Sc. degrees, both in electrical engineering from Cairo
University, Cairo, Egypt in 1963 and 1970, respectively, and his Ph. D. in electrical engineering from the University of Bradford,
Bradford, UK, in 1976. He joined the General Egyptian Aero Organization as an R & D engineer from 1963 to 1971, and the
Department of Electronics and Electrical Communications Engineering, Cairo University, since March 1971 till now, where he is
currently an emeritus professor. He was a research assistant with the Postgraduate School of Electrical Engineering, University
of Bradford, England from 1973 to 1976, and a visiting scholar at the Ecole National Superieure des Telecommunications,
Paris, in 1978. He was seconded as an associate professor to Qatar University, Electrical Engineering Department from 1982
to 1988, and as dean of Faculty of Computer Science and Information Systems, October 6 University from 1996 to 1998. His
areas of interest include speech and image signal processing, wireless communication systems, signal processing for
communications and software defined radio.
_________________________________________________________________________________
© Copyright (2022): Author(s). The licensee is the publisher (B P International).
DISCLAIMER
This chapter is an extended version of the article published by the same author(s) in the following journal.
Journal of International Journal of Computer and Communication Engineering, 3(1):2014.
139

Reem ICI Self Cancellation Book Chapter BP International

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

Reem ICI Self Cancellation Book Chapter BP International

Uploaded by

Copyright:

Available Formats

Chapter 9

Print ISBN: 978-93-5547-370-7, eBook ISBN: 978-93-5547-378-3

An Efficient ICI Self-cancellation Scheme for High

OFDM is an established modulation scheme in current broadband wireless mobile communication

Orthogonal Frequency Division Multiplexing (OFDM) is commonly considered to be the technical

Fig. 1. OFDM frequency spectrum [5]

where n is not equal to m.

1.1 Typical OFDM System

Fig. 2. The basic block diagram of an OFDM system in AWGN channel

1.1.1 Signal Mapping

1.1.2 Inverse Discrete Fourier Transform

1.1.3 Guard Interval and Cyclic Prefix

Fig. 3. Guard period insertion in OFDM

1.1.4 Communication Channel

1.2 Effect of Carrier Frequency Offset

In an AWGN channel, the frequency offset is modeled as a time-varying multiplicative factor

Fig. 4. Frequency Offset Model

The received signal is given as:

Fig. 5. The effect of CFO on the subcarriers orthogonally 9]

1.3 Factors Inducing ICI

1.3.1 Doppler Effect

f D  f D max cos (8)

1.3.2 Synchronization Errors

1.4 Methods of ICI Reduction

1.4.1 Time Domain Windowing

1.4.2 Frequency Domain Equalization

Fig. 6. Pilot subcarrier arrangement.

1.4.3 Pulse Shaping

1.4.4 Repeated Data Methods- ICI self-cancellation schemes

Fig. 7. Basic block diagram of an ICI self-cancellation OFDM systems

Table 1. 1 Different ICI Self- Cancellation Schemes

Schemes Description CIR Phase Rotation

CIRASR  Im[ 2S (0)  S (1)  S (1)]

Symmetric TX Signal Im[S (0)  S * (0)]

Symbol X*N/2+1, X*1, X*0 )

Schemes Description CIR Phase Rotation

General TX Signal Im[e  j S (0)  e j S * (0)]

2. PROPOSED ICI SELF -CANCELLATION SCHEMES

2.1 Single-Carrier Transmission

2.1.1 Conjugate Multiplicative Single-Carrier Transmission (CM-SC)

Eqn. (11) indicates that the CFO error is completely eliminated.

2.1.2 Conjugate Phase Additive Single-Carrier Transmission (CPA-SC)

We modified the above algorithm as follows:

2.2 MultiCarrier Transmission

2.2.1 Adjacent Conjugate Phase Additive OFDM (ACPA)

2.2.2 Symmetric Conjugate Phase Additive OFDM (SCPA)

Input Binary Modulator

Fig. 8. Proposed ACPA OFDM System model.

Fig. 9. Proposed SCPA OFDM System model

3. SIMULATION RESULTS AND PERFORMANCE EVALUATION

 Graceful degradation for high ε values.

3.1 Conjugate Phase Additive Single Carrier Transmission (CPA-SC)

3.1.1 CPA-SC 16-QAM System

3.1.2 CPA-SC 256-QAM System

3.2 Conjugate Phase Additive OFDM Transmission

3.2.1.1 CIR Performance

CIR versus  for OFDM systems

Fig. 12. The CIR performance of the ACPA, SCPA schemes

CIR versus  for OFDM systems

Fig. 13. CIR improvement using different ICI self-cancellation schemes

3.2.1.2 Phase Shift

3.3 BER Performance Comparison with Other Schemes

3.3.1 QPSK system

Symbol XN/2+1, X1, X*0 )