0% found this document useful (0 votes)
92 views

Performance and Evaluation of Ofdm and Sc-Fde Over An Awgn Propagation Channel Under RF Impairments Using Simulink at 60Ghz

antnne

Uploaded by

elhoucine
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
92 views

Performance and Evaluation of Ofdm and Sc-Fde Over An Awgn Propagation Channel Under RF Impairments Using Simulink at 60Ghz

antnne

Uploaded by

elhoucine
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

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

net/publication/266592165

Performance and evaluation of OFDM and SC-FDE over an AWGN propagation


channel under RF impairments using simulink at 60GHz

Conference Paper · November 2014


DOI: 10.1109/LAPC.2014.6996487

CITATIONS READS

13 1,336

6 authors, including:

Rodolfo Gomes Zaid Al-Daher


TWEVO Technologies University of South Wales
17 PUBLICATIONS   50 CITATIONS    12 PUBLICATIONS   55 CITATIONS   

SEE PROFILE SEE PROFILE

A. Hammoudeh Khaled Sobaihi


University of South Wales DST Innovations Ltd
67 PUBLICATIONS   390 CITATIONS    12 PUBLICATIONS   44 CITATIONS   

SEE PROFILE SEE PROFILE

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

MMSuper View project

Hybrid FSS and rectenna design for wireless power harvesting View project

All content following this page was uploaded by Rafael F. S. Caldeirinha on 23 July 2015.

The user has requested enhancement of the downloaded file.


2014 Loughborough Antennas and Propagation Conference (LAPC) 10 - 11 November 2014, UK

Performance and Evaluation of OFDM and SC -


FDE over an AWGN Propagation channel under RF
impairments using Simulink at 60GHz
Rodolfo Gomes1,2 , Zaid Al-Daher1 , Akram Hammoudeh1 , Khaled Sobaihi1 , Rafael Caldeirinha1,2 , Telmo Fernandes1,2
1
Faculty of Computing, Science and Engineering, University of South Wales, United Kingdom
2
Instituto de Telecomunicações (DL-IT), ESTG, Polytechnic Institute of Leiria, Leiria, Portugal
[email protected]

Abstract—This paper presents an implementation of the IEEE Moreover, the proposed Simulink based framework can be used
802.15.3c standard, in Simulink, for high data-rate applications. to provide studies on the impact of radio channel impairments,
The recent standard IEEE 802.15.3c provides the implementa- e.g. under severe multipath conditions, in the performance of
tion of such wireless communication system using the 60GHz
frequency band. OFDM and SC-FDE are the two transmis- a communication link at 60GHz. And thus, the combination of
sion schemes adopted by the standard. Performance of these both RF non-idealities and propagation channels impairments
technologies at mm-wave signals are severely affected by non- are essential for appropriate RF system design and modelling.
linearity of the RF front-ends. This paper presents the impact
of RF impairments for each transmission scheme along with
II. IEEE S TANDARD S PECIFICATIONS
comparative analysis, in terms of BER and PSNR.
Index Terms—OFDM, SC-FDE, LDPC, IEEE 802.15.3c, The recent IEEE 802.15.3c standard [4] describes the phys-
60GHz.
ical layer details of communication systems operating at the
60GHz frequency band, with date rates between 2Gb/s and
I. I NTRODUCTION 5Gb/s depending on the propagation scenario and the purpose
Orthogonal Frequency Division Multiplexing (OFDM) and of the radio link. OFDM and SC-FDE are the transmission
Single Carrier with Frequency Domain Equalization (SC-FDE) schemes adopted by this standard.
transmission schemes are seen as the two alternatives to The main PHY operating parameters considered for OFDM
provide high-data rate links. The baseband signal is upconver- and SC block transmission are given in Table I.
ted to the passband by In-phase and Quadrature (IQ) mixing
and is then transmitted over the radio channel, including the Table I
antenna. In the receiver side, the signal is down converted S UMMARY OF THE MAIN PARAMETERS CONSIDERED IN THE DESIGN OF
THE OFDM AND SC-FDE MODELS .
to baseband, by using another mixer. The Radio frequency
(RF) front-ends at 60GHz induce several critical non-idealities Parameter Value
which affects both transmitted and received radio signals. Non- FFT size block (Nf f t ) 512
Cyclic prefix length (Ncp ) 64 samples
idealities must be taken into account in the system design of FFT sampling rate 250MHz
both schemes. These non-linearities include the effect of the Data rate 1Gbps
phase noise (PN), IQ imbalances and power amplifier (PA) Modulation 16 QAM
FEC code LDPC(672,336)
non-linear distortions. Code rate 1/2
The literature contains a limited number of articles directly
related in addressing the RF impairments at 60GHz. For
example, articles [1]–[3] compare both schemes, however the
III. L OW D ENSITY PARITY C HECK CODES
type of modulation used in both is not the same as the
one(s) suggested by the standard [4]. In other papers the Low Density Parity Check codes (LDPC) codes make part
RF impairment models used are not appropriate for 60GHz, of the linear block codes family. Unlike other linear block
according to [5]. codes, the LDPC codes are not defined via the generator matrix
In this paper, SC-FDE and OFDM transmission schemes G, but rather via a parity matrix H. Therefore, there is a higher
are implemented in Simulink [6], [7] according to the IEEE complexity with the encoding process than with the decoding
802.15.3c standard. For both systems, the BER is calculated one. LDPC codes are decoded by the Sum-Product Algorithm
over an Additive White Gaussian Noise (AWGN) channel in (SPA) iteratively. To implement this type of codes in Simulink,
order to evaluate their performance in an ideal communication it is necessary to insert the matrix H presented in [4] for a code
channel under the effects of the RF frontend, at 60GHz. rate of 1/2, to the encoder and decoder blocks.

978-1-4799-3662-5/14/$31.00 ©2014 IEEE 685


IV. S IMULINK S YSTEM M ODELING
OFDM is the transmission of N parallel data symbols, where
each one is assigned to a different subcarrier with Quadrature
Amplitude Modulation (QAM). To transmit a certain data at
a specified rate, the baudrate per each subcarrier is as slower
as the total number of subcarriers in the system, increasing
the OFDM symbol time. This fact leads to the mitigation of
the negative effects on a wireless system due to a frequency
selective radio channel (reduction of the channel equalization
complexity).
Let Xn , n = 0,1,..., N-1, be the N data symbols to be
transmitted over the N subcarriers, where Xn is the complex Figure 2. OFDM Simulink block diagram.
sequence corresponding to a two-dimensional QAM constel-
lation interpreted in the frequency domain.
The transmitted baseband OFDM signal can be written as
[8]:
N −1

x(mTs ) = Xn ej2πnm/N (1)
n=0

From (1), it can be seen that an OFDM symbol is generated


in discrete time using the IFFT, which modulates the data
symbols. In the receiver, the Fast Fourier Transform (FFT) is
used to demodulate the data symbols due to the orthogonality Figure 3. SC FDE Simulink block diagram.
properties between the IFFT and FFT.
SC - FDE can be seen as the bridge between SC and OFDM. system using LDPC(672,336) forward error correction code
In Fig. 1 a conversion from OFDM to SC-FDE or SC - FDE is also presented. It is noticed that FEC code improves the
to OFDM, is presented. performance of both systems by about 8 to 9 dB.

Transmitter

IFFT CP insertion

IFFT is switched to transmitter for OFDM


or switched to receiver for SC - FDE
Receiver

Detector IFFT Equalizer FFT

Figure 1. Conversion between OFDM and SC-FDE.

From Fig. 1, it can be observed that the main difference


between both systems mainly rely on the position of the IFFT (a) (b)
block. In an OFDM system, the IFFT blocks is placed at Figure 4. a) Coded OFDM vs uncoded BER and b) Coded SC - FDE vs
the transmitter in order to allocate data symbols into different uncoded BER
subcarriers. However, in a SC-FDE system, such block is at the
In this study, the number of bits per Simulink iteration and
receiver side to transform the equalized data into time domain
the maximum number of error bits per iteration used were 10,
for signal detection. The overall complexity of both systems
8 and 200, respectively. The number of iterations used in the
is, however, similar yielding to the same length of data blocks.
LDPC decoder block in order to converge the algorithm were
Figs. 2 and 3, show the Matlab Simulink diagram block of
20.
both transmission schemes based on Fig.1.
V. RF I MPAIRMENT MODELS FOR 60 GH Z AND ITS
A. Modelling Validation
EFFECTS ON THE OFDM AND SC-FDE SYSTEM
The BER is obtained for each iteration of the energy per PERFORMANCE
bit to noise power spectral density ratio (Eb /No ). Figs. 4a
and 4b show the comparison between the theoretical BER A. Phase Noise from Oscillators
curves of both uncoded systems in comparison with simu- In a typical communication system, mixers are used in the
lations. The BER performance of the OFDM and SC-FDE upconverter and downconverter stages, in the transmitter and

686
receiver side, respectively. However, if the oscillator used to B. IQ imbalances
feed the mixer placed in the receiver side of the system, has
IQ imbalances are characterized by the phase and amplitude
a different carrier frequency, inter-carrier interference (ICI)
offset between the I and Q branches in quadrature modu-
and common phase rotation (CPE) in a OFDM system and
lators/demodulators. It is important to include I/Q mismatch
phase rotation and intersymbol interference (ISI) for a SC-
in any system simulation for precise performance simulation.
FDE system will occur.
According to [8], the complex baseband output signal, which
Considering it is impossible to have two oscillators oper- suffer from I/Q mismatches can be expressed by:
ating at the same carrier frequency at mm-wave band, the
distortion in the OFDM signal and in the SC-FDE due to
this effect is always present. This instability of carrier signal y(t) = (cos(Δφ) + jsin(Δφ)).x(t)+,
generators used at the transmitter and receiver is known as PN
and according to [5] can be simply modelled as an high-pass (.cos(Δφ) − jsin(Δφ)).x(t)∗ , (3)
filter, which yields one-pole and one-zero:
where, x(t) is the input signal, x(t)∗ is its complex conjug-
2
[1 + (f /fz ) ] ate,  and Δφ are the gain and phase mismatches.
P SD(f ) = P SD(0). , (2) From (3), is it seen that the desired signal x(t) is interfered
[1 + (f /fp )2 ]
by its image signal, x(t)* , causing attenuation and a rotation
where, P SD(0) is the low frequency phase noise, fp is the of the desired signal. In an OFDM system ICI will occur and
pole frequency and fz is the zero frequency. in a SC FDE, a complex conjugated symbol will be added,
Fig. 5, shows the representation of the 60GHz phase noise which causes ISI [1].
model with fp = 1M Hz and fz = 100M Hz. Figs. 7a, 7b, demonstrate the effect of IQ imbalances in the
OFDM system, and Figs. 8a and 8b show the same effect for
the SC-FDE system.

Figure 5. Phase noise model vs. PSD(0) constants. (a) (b)


Figure 7. OFDM BER performance vs. IQ imbalances: a) Gain mismatch b)
Figs. 6a and 6b show the effect of the phase noise on the Phase mismatch.
performance of both transmission schemes.

(a) (b)
(a) (b)
Figure 8. SC-FDE BER performance vs IQ imbalances: a) Gain mismatch
Figure 6. BER performance vs. phase noise effect: a) OFDM b) SC-FDE. b) Phase mismatch.

Analyzing both figures, it can be seen that the degradation Comparing Figs. 7 and 8, it is clear that the single carrier
performance due to phase noise is very similar in both systems, system is slightly more robust to the IQ imbalances than
although OFDM slightly outperforms the SC system. OFDM.

687
C. Power Amplifier
G.|x[n]|
Power amplifiers are an essential component of a wireless FAM −AM (y) = 1 , (6)
communication system. However, they introduce distortion in (1 + ( |x[n]| 2p 2p
Vsat ) )
the signal bandwidth and outside the power spectrum of the A.|x[n]|q
transmitting signal. To avoid this distortion the amplifier must FAM −P M (y) = , (7)
(1 + ( |x[n]|
B ) )
q
be set to operate in its linear zone. For instance, if the PA starts
working outside its linear zone harmonics or intermodulation respectively, where G is the gain of the PA, Vsat is the
products will appear in the spectrum of the signal, contributing saturation voltage level, |x[n]| is the input voltage level, p
to the degradation of the signal. In addiction, if the envelope is the smoothness factor, which controls the curvature of the
of the input signal is not constant the operation point of the PA saturation transfer function of the PA and A, B and q are
will change. Therefore in this case its power amplitude output fitting parameters.
should be backed off to keep the signal within the linear zone Note: p = 1.1, Vsat = 1.9 , A = −345, B = 0.17 and
at the expense of less power efficiency. q = 4 are the values used by the IEEE 802.13.3c 60 WPAN
The PA output back-off (OBO) is given by: standardization group for PA impact evaluation [10].
The effect of the modelling of the IEEE standard PA model
Psat for OFDM and SC-FDE is presented in Figs. 10 and 11,
OBO = 10.log10( ), (4)
Pout consequently.
where, Pout is the average power emitting from a PA and
Psat is the saturation output power.
The parameter OBO is more important in an OFDM system
than in a SC-FDE system, since its signal is characterized to
have large amplitude fluctuations, which make it susceptible
to system non-linearities. The amplitudes fluctuation are char-
acterized by the peak to average power ratio (PAPR) of the
signal. The PAPR for the discrete-time signal is expressed as:

|x[n]|2
P AP R(x[n]) = max0≤n≤N −1 , (5)
E[|x[n]|2 ]
(a) (b)
where, |x[n]|2 is the maximum instantaneous power, E[.] is Figure 10. OFDM under PA non-linearities: a) BER b)Signal degradation.
the average power and N is the number of the subcarrier per
OFDM signal.
Figures 9a and 9b, show the simulated PAPR distribution
over 10000 samples of the OFDM signal and its cumulative
distribution function.

(a) (b)
Figure 11. SC-FDE under PA non-linearities a) BER b)Signal degradation.

By comparing the above figures, it can be seen that SC-FDE


(a) (b) is more robust against such non-linearities. i.e to ensure a BER
Figure 9. Simulated OFDM PAPR representation: a) Samples b) Cumulative of 10−6 for high data rates transmissions, the SC system has,
distribution function. approximately, 2.7dB of losses for IBO= 13. However, OFDM
presents 3dB of losses for a input back-off (IBO)= 15.
From these figures the simulated OFDM signal has a PAPR Note: The OBO and IBO in the Simulink system represents
 9.3dB and 90% of the samples of the signal have a PAPR the same, since the PA model has a gain of 0dB.
lower than 3.6dB.
According to [5], by using the modified Rapp model, the VI. E FFECT OF THE PA NON - LINEARITIES ON THE VIDEO
amplitude modulation to amplitude modulation (AM-AM) [9] QUALITY FOR BOTH SCHEMES
and amplitude modulation to phase modulation (AM-PM) In order to asses the quality evaluation of an uncompressed
effects of 60GHz PA can be modeled by: video content transmitted, using both transmission schemes,

688
over an AWGN channel under power amplifier non-linear
distortions, the peak signal-to-noise ratio (PSNR) metric is
used.
The AM/AM PA model presented in the section V-C,
with the parameters set according to the IEEE 802.13.3c
60GHz WPAN standardization group, is implemented and
an uncompressed video frame with XVGA(1024x768.yuv)
format, shown in Fig 12, is transmitted.
(a) (b)
Figure 14. PA non-linearities impact on the subjective quality of the
transmitted frame for a Eb/N0 = 16: a) SC FDE b) OFDM.

The simulation results show that SC-FDE clearly outper-


forms OFDM under the PA effect and the signal degradation
is reduced for low PA IBO. OFDM being characterized for
having a higher PAPR value than SC-FDE, degradations
are more pronounced. However, under phase noise and IQ
imbalances the performance is similar and for the phase noise
Figure 12. Reference frame from the Kendo video sequence for the PSNR
effect, OFDM presents a slightly better performance.
calculation. As for the PSNR analysis, it can be concluded that the
SC-FDE ensures a much better quality on the video than
Fig. 13 contains the results obtained by applying the PSNR
OFDM. Therefore, the use of FEC codes is indispensable to
in the received frame for different channel conditions and
increase the video quality in a OFDM transmission, i.e. the
under the non-linear PA effect.
LDPC(672,336) improves in the system performance by about
8 or 9dB.

R EFERENCES
[1] F. Horlin and A. Bourdoux, “Comparison of the sensitivity of OFDM
and SC-FDE to CFO, SCO and IQ imbalance,” in 3rd International
Symposium on Communications, Control and Signal Processing, 2008.
ISCCSP 2008, Mar. 2008, pp. 111–116.
[2] M. Lei, I. Lakkis, C.-S. Sum, T. Baykas, J.-Y. Wang, M. Rahman,
R. Kimura, R. Funada, Y. Shoji, H. Harada, and S. Kato, “Hardware
impairments on LDPC coded SC-FDE and OFDM in multi-gbps WPAN
(IEEE 802.15.3c),” in IEEE Wireless Communications and Networking
Conference, 2008. WCNC 2008, Mar. 2008, pp. 442–446.
Figure 13. PA non-linearities impact in the PSNR of the transmitted frame [3] U. Rizvi, G. Janssen, and J. Weber, “Impact of RF impairments
for different channel conditions and transmission schemes. on the performance of multi-carrier and single-carrier based 60 GHz
transceivers,” in 2007 14th IEEE Symposium on Communications and
Vehicular Technology in the Benelux, Nov. 2007, pp. 1–5.
Analysing Fig. 13, without the effect of the non-linearities [4] “802.15.3c-2009 - physical layer(PHY) specifications for high rate
from the PA, the quality of the video increase as the Eb/N 0 in- wireless personal area networks(WPANs).”
creases as expected. Clearly the effect of these non-linearities [5] Yozo Shoji, Hiroshi Harada, Ryuhei Funada, Shuzo Kato, Kenichi
Maruhashi, Ichihiko Toyoda, and Kazuaki Takahashi, “RF impairment
have a different impact in the video quality, depending on the models for 60ghz-band SYS/PHY simulation.”
transmission scheme. For the OFDM, it can be seen that the [6] T. P. Surekha, T. Ananthapadmanabha, and C. Puttamadappa, “Mod-
PSNR remains constant as Eb/N 0 increases at approximately eling and performance analysis of QAM-OFDM system with AWGN
channel,” in 2011 Third Pacific-Asia Conference on Circuits, Communi-
9dB, unlike the SC-FDE scheme. Therefore, SC-FDE scheme cations and System (PACCS), Jul. 2011, pp. 1–4.
ensures a better quality of the transmitted video over severe [7] J. Zhang and Z. Zhang, “Simulation and analysis of OFDM system based
channel conditions and non-linearities. on simulink,” in 2010 International Conference on Communications,
Circuits and Systems (ICCCAS), Jul. 2010, pp. 28–31.
Fig. 14, shows the subjective quality of a received frame [8] Su-Khiong (SK) Yong, Pengfei Xia, and Alberto Valdes Garcia, 60GHz
from the video sequence transmitted for the both schemes un- TECHNOLOGY For Gbps. Wiley.
der the effect of distortions from the PA, for Eb/N 0 = 16dB. [9] Christoph Rapp, “Effects of HPA-nonlinearity on a 4-DPSK/OFDM-
signal for a digital sound broadcasting system,” 1991.
VII. C ONCLUSIONS [10] A. Maltsev, A. Lomayev, A. Khoryaev, A. Sevastyanov, and R. Maslen-
nikov, “Comparison of power amplifier non-linearity impact on 60
In this paper, a performance comparison of OFDM and SC- GHz single carrier and OFDM systems,” in 2010 7th IEEE Consumer
FDE as transmission schemes for 60GHz systems, under RF Communications and Networking Conference (CCNC), Jan. 2010, pp.
circuit imperfections is presented. 1–5.

View publication stats


689

You might also like