(issn:2348-0246(online)),volume-1,issue-1,january ...ijetts.org/admin/issues/ijetts-010104.pdf ·...

INTERNATIONAL JOURNAL OF EMERGING TRENDS IN TECHNOLOGY AND SCIENCES(ISSN: 2348-0246(online)), VOLUME-1, ISSUE-1, JANUARY-2014

29www.ijetts.org

OFDM SYSTEMS: ICI SELF CANCELLATION FOROPTIMIZED DATA ALLOCATION

Srinivas Surishetty1, R.Sathish Kumar2

Abstract—Symmetric symbol repeat (SSR) intercarrierinterference (ICI) self cancellation scheme has proved tobe a simple and convenient technique to reduce ICIcaused by frequency offsets. It utilizes data allocation andcombining of (1,-1) on two symmetrically placedsubcarriers to mitigate the effect of ICI. However, the dataallocation factors (1,-1) are not an optimum. In thispaper,an optimum data allocation and Combining schemeis proposed to maximize CIR performance for anestimated normalized frequency offset. But, this requirescontinuous CFO estimation and feedback circuitry. A

sub-optimal scheme utilizing sub-optimal pair ( so, so) is alsoproposed to completely eliminate the requirement of CFOestimation.Simulation results confirm the outperformanceof the proposed optimal scheme over conventional SSRICI self cancellation scheme. Sub-optimal scheme can beapplied for the any range of and a sub optimum value canbe calculated using proposed sub-optimal scheme. TheCIR of SSR ICI self cancellation scheme using theproposed sub-optimal approach is also found to be betterthan conventional SSR ICI self cancellation.

1. INTRODUCTION

In recent years, direct conversion receiver hasdrawn a lot of attention due to its low powerconsumption and low implementation cost. Howeversome mismatches in direct conversion receiver canseriously degrade the system performance, such asin-phase and quadrature-phase (I/Q) imbalance andcarrier frequency offset (CFO). The I/Q imbalance isdue to the amplitude and phase mismatches betweenthe I and Q-branch of the local oscillator whereas theCFO is due to the mismatch of carrier frequency atthe transmitter and receiver. It is known that the I/Qimbalance and CFO can cause a serious inter-carrierinterference (ICI) in orthogonal frequency divisionmultiplexing (OFDM) systems. As a result, the biterror rate (BER) has an error-flooring. In assumingthat the channel frequency response is smooth, afrequency-domain estimation method has beenproposed to jointly estimate the I/Q imbalance andchannel frequency response. Recently, exploiting thefact that the size of the DFT matrix is usually largerthan the channel length in OFDM systems, atime-domain method was proposed for the jointestimation of I/Q and channel response. Both thefrequency-domain and time-domain methods needonly one OFDM block for training and can achieve agood performance. For CFO estimation, a low

complexity maximum likelihood (ML) technique wasproposed

The joint estimation of CFO and I/Q imbalance wasinvestigated .The authors combined the techniques tojointly estimate the I/Q and CFO parameters usingtwo repeated training blocks. The Cramer-Rao boundof CFO

1.1 Digital Communication Systems

A digital communication system is often dividedinto several functional units. The task of the sourceencoder is to represent the digital or analoginformation by bits in an efficient way. The bits arethen fed into the channel encoder, which adds bits ina structured way to enable detection and correction oftransmission errors. The bits from the encoder aregrouped and transformed to certain symbols, orwaveforms by the modulator and waveforms aremixed with a carrier to get a signal suitable to betransmitted through the channel. At the receiver thereverse function takes place. The received signals aredemodulated and soft or hard values of thecorresponding bits are passed to the decoder. Thedecoder analyzes the structure of received bit patternand tries to detect or correct errors. Finally, thecorrected bits are fed to the source decoder that isused to reconstruct the analog speech signal or digitaldata input. The main question is how to designcertain parts of the modulator and demodulator toachieve efficient and robust transmission through amobile wireless channel. The wireless channel hassome properties that make the design especiallychallenging: it introduces time varying echoes andphase shifts as well as a time varying attenuation ofthe amplitude

. For mobile or wireless applications,the channel is often described as a set of independentmultipath components. Among the most importantparameters when choosing the modulation schemeare the delay and the expected received power fordifferent delays. Large delays for stronger pathsmean that the interference between the differentreceived signal parts can severe, especially when thesymbol rate is high so that the delay exceeds several


30www.ijetts.org

symbols. In that case one has to introduce anequalizer to mitigate the effects of inter-symbolinterference (ISI). Another alternative is to use manyparallel channels so that the symbol time on each ofthe channels is long. This means that only a smallpart of the symbol is affected by ISI and this is theidea behind orthogonal frequency divisionmultiplexing.

1.2 Evolution of Telecommunication Systems

Many mobile radio standards havebeen developed for wireless systems throughout theworld, with more standard likely to emerge. Mostfirst generations systems were introduced in the mid1980s, and can be characterized by the use of analogtransmission techniques, and the use of simplemultiple access techniques such as FrequencyDivision Multiple Access (FDMA). First generationtelecommunications systems such as AdvancedMobile Phone Service (AMPS), only provided voicecommunications. They also suffered from a low usercapacity, and security problems due to the simpleradio interface used.

Second generation systems were introducedin the early 1990s, and all use digital technology.This provided an increase in the user capacity ofaround three times. This was achieved bycompressing the voice waveforms beforetransmission.

Third generation systems are an extensionon the complexity of second generation systems andare already introduced. The system capacity isexpected to be increased to over ten times originalfirst generation systems. This is going to be achievedby using complex multiple access techniques such asCode Division Multiple Access (CDMA), or anextension of TDMA, and by improving flexibility ofservices available.

because each chip of the PN sequence experiencesindependent fading, which tends to destroy theorthogonality between spreading sequences. Thisincreases the MAI and degrades the BERperformance. As a result, MC-CDMA systemsperform best under low channel loads.

2. FUNDAMENTALS OF OFDM

In this chapter, the basic principles ofOFDM systems are discussed based on a basic

system model of OFDM modulation. In addition tothis the problems that make the implementationdifficult are discussed and analyzed.

2.1 Basic principles of OFDM systems

2.1.1 Introduction

OFDM is a multi channel modulatingtechnique that makes use of Frequency DivisionMultiplexing (FDM) of orthogonal multi carriersbeing modulated by a low bit rate digital stream. InFDM, inter channel interference is diminished by thedeterrence of the spectral overlapping of sub-carriersbut it guides to an inadequate use of spectrum. Toprevail over this obstacle OFDM uses orthogonalsub-carrier that helps an efficient use of the spectrum.

· Emitter and receiver are efficientlyimplemented with IFFT/FFT

· Throughput maximization· Effectiveness against channel distortion

Multi-path delay spread toleranceEven though OFDM has numerous advantages it

has some disadvantages on the quality of theanalogue radio frequency front end of bothtransmitter and receiver:

· Sensitiveness to carrier frequency errors· To maintain the orthogonality between

subcarriers, the amplifiers need to be linear· OFDM systems have high peak-to average

ratio which may require large amplifier.

2.1.2 Generation of OFDM signals

To put in practice the OFDM transmissionmethod, the message signal has to be digitallymodulated. To maintain orthogonality between thesub carriers the carrier is split into lower frequencysub carriers.

2.1.3 Transmitter

Modulating schemes such as Quadratureamplitude modulation (QAM) (16QAM or 64QAMfor example) or some form of phase shift keying(PSK) are used to modulate the message signal.Before passing to the IFFT block the modulatedsignal has to be converted to parallel signals. TheIFFT operation will take place to form a singlecombined signal that can be transmitted. Thesecombined signals are orthogonal subcarriers.


31www.ijetts.org

Fig1: An OFDM system in Frequency Domain

2.1.4 Guard period

When OFDM signals aretransmitted multipath delay could affect thetransmission. The inter symbol interference (ISI)caused by multipath propagation can be avoided byinserting guard interval between every OFDMsymbol. The insertion of the guard period betweentransmitted symbols is vital if the signals insubcarriers are to hold on to the orthogonality duringtransmission process. The guard period allowsmultipath components from previous symbol to dieout before information from the next OFDM symbolis recorded.

Fig2. Transmitter and Receiver Block of an OFDMsystem.

2.1.5 Channel model

Transmitted signals are affected by multipathpropagation. These signals are reflected, refracted orscattered by objects that come across duringtransmission before it reaches the intended receiver.Same signals appear at the receiver with differentphase, amplitude and with different time delays.

These signals are added destructively at the receiverand make the received signal to fade. Multipathfading environment and block fading channel model(the fading gain process is piecewise constant onblocks of N symbols and changes to otherindependent value) is assumed. Figure shows thechannel is represented by FIR system for one datapacket .

Fig3. channel representation

Where: Yk is the received signal,is the channel matrix, xk is the input symbol and vkis the additive noise.

3. MIMO OFDM SYSTEMS

3.1 MIMO System

MIMO signaling is a groundbreakingdevelopment pioneered by Jack Winters of BellLaboratories in his 1984 article [2]. Several differentantenna configurations are used in definingspace-time systems.

3.1.1 Basic Structure of MIMO System

There exist several communication transmissionmodels as follows (see Fig.3):

1. Single-input-and-single-output (SISO) system: It isuses only one antenna both at the transmitter andreceiver.

2. Single-input-and-multiple-output (SIMO) system:It uses a single transmitting antenna and multiplereceiving antennas [3].

3. Multiple-input-and-single-output (MISO) system:It has multiple transmitting antennas and onereceiving antenna.

4. Multiple-input-multiple-output (MIMO) system: Ituses multiple antennas both for transmission andreception. Multiple transmitting and receivingantennas will achieve antenna diversity withoutreducing the spectral efficiency.


32www.ijetts.org

Fig4. Multi-transmitting and receiving antennas

3.2 MIMO-OFDM System

In high-speed wireless communication,combining MIMO and OFDM technology, OFDMcan be applied to transform frequency-selectiveMIMO channel into parallel flat MIMO channel,reducing the complexity of the receiver, throughmultipath fading environment can also achieve highdata rate robust transmission. Therefore,MIMO-OFDM systems obtain diversity gain andcoding gain by space-time coding, at the same time,the OFDM system can be realized with simplestructure. Therefore, MIMO-OFDM system hasbecome a welcome proposal for 4G mobilecommunication systems.

3.2.1 Basic Structure of MIMO-OFDM System

At the transmitting end, anumber of transmission antennas are used. An inputdata bit stream is supplied into space-time coding,then modulated by OFDM and finally fed to antennasfor sending out (radiation). At the receiving end,in-coming signals are fed into a signal detector andprocessed before recovery of the original signal ismade. Fig.5 shows the basic structure of aMIMO-OFDM system.

Figure5 Basic structure of MIMO-OFDMsystem.

Presently, many companies and research institutionshave developed MIMO-OFDM experimental systems.Airbust –production of Iospan Company that firstused MIMO and OFDM technology in the physicallayer at the same time for wireless communicationsystems.

4. I/Q IMBALANCE

OFDM systems are susceptible to analogfront end imperfections; IQ imbalance is one of theimpairments that cause the received symbols not tobe correctly demodulated. The thesis is all aboutreduction of this problem in OFDM systems. Becauseof this front end imperfection the analog part is theexpensive one in the system. Down conversion is abasic phase in all radio frequency (RF) architectures.To convert RF signal into its equivalent basebandmany configuration are used. The traditional OFDMsystem employs the Superhetrodyne receiver toconvert RF signal into baseband and vice versa. Thisarchitecture use intermediate frequencies (IF) toconvert RF signal down to baseband. Thedisadvantage of this configuration is it uses a lot ofcomponents which are expensive and external to geta good signal quality.

4.1 I/Q mismatch

This document describes the impact of amplitudeand phase mismatch in an I/Q(in-phase/quadrature-phase) signal pair. The result ofa mismatch is an unwanted spectral component at thenegative signal frequency. An ideal sinusoidal I/Qsignal pair with no amplitude or phase mismatch canbe represented by the following complex function:

Equation 1:

y(t) = A exp(i(ωt + φ)) = A cos(ωt + φ) + i Asin(ωt + φ)

The real part (the first term in the second expression)of y(t) is the I signal, and the imaginary part (thesecond term except the "i") is the Q signal. ω, A, andφ are the angular frequency, amplitude and phase,respectively.

If an amplitude and phase mismatch is introduced,the function becomes:

Equation 2:

y(t) = A cos(ωt + φ) + i α A sin(ωt + φ + ε)


33www.ijetts.org

Here, A and φ are the amplitude and phase of the Isignal, α is the amplitude mismatch (the ratio of Qamplitude to I amplitude), and ε is the phasemismatch (the Q phase plus 90 degrees minus the Iphase). For α=0 dB (α=1) and ε=0 degrees there is nomismatch.

By rewriting Equation 2 we can split the mismatchedI/Q signal into its spectral components:

Equation 3:

y(t) = 0.5 (1 + α exp(i ε)) A exp(i(ωt + φ)) + 0.5(1 - α exp(-i ε)) A exp(-i(ωt + φ))

The first term is the wanted signal (including both Iand Q parts) - a pure complex sinusoid at thefrequency ω. The second term is an unwanted signalat the negative frequency -ω. The wanted signal issimply the ideal I/Q signal from Equation 1multiplied by a complex constant, which modifies theamplitude and phase (here, the I and Q parts aremodified equally).

4.2 Mean Square Error

In statistics, the mean square error orMSE of an estimator is one of many ways to quantifythe difference between an estimator and the truevalue of the quantity being estimated. MSE is a riskfunction, corresponding to the expected value of thesquared error loss or quadratic loss. MSE measuresthe average of the square of the "error." The error isthe amount by which the estimator differs from thequantity to be estimated. The difference occursbecause of randomness or because the estimatordoesn't account for information that could produce amore

The MSE of an estimator with respect to theestimated parameter θ is defined as

The MSE is equal to the sum of the variance and thesquared bias of the estimator

The MSE thus assesses the quality of an estimator interms of its variation and un biasedness. Note that theMSE is not equivalent to the expected value of theabsolute error.

4.3 Carrier Frequency Offset Estimation

OFDM is a great technique to handle impairments ofwireless communication channels such as multipathpropagation. Hence, OFDM is a practical candidatefor future 4G wireless communications techniques .On the other hand, one of the major drawbacks of theOFDM communication system is the drift inreference carrier. The offset present in receivedcarrier will lose orthogonality among the carriers asshown in Figure. Hence, the CFO causes a reductionof desired signal amplitude in the output decisionvariable and introduces ICI. Then it brings up anincrease of BER. The effect caused by CFO forOFDM system was analyzed BER upper bound ofOFDM system is analyzed without ICI selfcancellation and BER of OFDM system is analyzedusing self cancellation.,. Thus, carrier frequencyoffset greatly degrades system performance.Therefore, practical OFDM systems need the CFO tobe compensated with sufficient accuracy, and this hasled to a whole lot of literature on CFO estimationalgorithms

Fig6. Schematic picture of the orthogonalsubcarriers in the existence of the CFO.

5. RESULTS& DISCUSSION

In this paper, we have considered an OFDMsystem with N 256 subcarriers and QPSK modulationscheme is used to modulate each of the subcarriers.The simulation model of the OFDM system is shownin Fig.1. The computer simulation using MATLABare performed to evaluate CIR and BERperformance. Fig. 2 shows the CIR performance ofstandard OFDM system, SSR ICI self-cancellation[6], Proposed SSR ICI self cancellation using optimal& sub-optimal approach. Fig. 3 shows BERperformance of the standard OFDM system,conventional SSR ICI self cancellation and theproposed SSR ICI self cancellation using sub-optimalapproach.

As seen from Fig. 2 the CIR performance ofthe proposed optimal approach is about 20dB betterthan the conventional SSR ICI self cancellationscheme. However, the proposed sub-optimalapproach also provides better CIR scheme


34www.ijetts.org

performance over conventional SSR ICI selfcancellation scheme, proposed suboptimal approachprovides a gain of more than 10dB at 0.15 overconventional SSR ICI self cancellation scheme. TheCIR performance of proposed SSR ICI selfcancellation scheme is slightly worse thanconventional SSR ICI self Cancellation scheme for[0.03,0.08]. The BER performance of the proposedSSRICI self cancellation scheme is very much improvedin comparison to standard OFDM system and veryclose to conventional SSR ICI self cancellationscheme.

Comparison of CIR Performance90

Standard OFDM system80 SSR ICI Self Cancellation

70Proposed SSR ICI Self Cancellation(Optimal)Proposed SSR ICI Self Cancellation(Sub-optimal)

60

CIR(dB) 50

40

30

20

10

00.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.240.04

Normalized frequency offset

Figure 2. CIR performance comparison ofvarious ICI self cancellation scheme

BER Comparison for epsilon=0.15100

10-1

10-2

BER

10-3

10-4Standard OFDM SystemConventional SSR ICI Self CancellationProposed SSR ICI Self Cancellation(Sub Optimal)

0 2 4 6 8 10 12Eb/No(dB)

Figure 3. BER performance comparison

7. REFERENCES

[1] Y.Wu andW. Y. Zou, “Orthogonal frequency divisionmultiplexing: A multicarrier modualtion scheme”, IEEETransaction on Consumer Electronics, vol. 41, no. 3,pp392-399, Aug. 1995.

[2] P.H.Moose, “ A Technique for orthogonal frequency divisionmultiple - xing frequency offset correction,’’IEEE Trans.Commun., vol.42, pp 2908-2914, Oct. 1994.

[3] Y. Zhao and S.G. Haggman, “Intercarrier InterferenceSelf-Cancellation Scheme for OFDM MobileCommunication Systems’’,IEEE Trans.Commun., vol. 49,no. 7, pp. 1185-1191, Jul. 2001.

[4] H.-G. Yeh, Y.-K. Chang, and B. Hassibi, “A scheme forcancelling intercarrier interference using conjugatetransmission in multicarrier communication systems," IEEETrans. Wireless Communication., vol. 6, no. 1, pp. 3-7, Jan.2007.

[5] Chin-Liang Wang and Yu-Chih Huang, “IntercarrierInterference Cancellation Using General Phase RotatedConjugate Transmission for OFDM Systems” IEEETransactions on Communications, vol. 58, no. 3, pp.812-819, March 2010.

[6] K.Sathanantham , R.M.A.P.Rajatheva and Slimane BenSlimane “Analysis of OFDM in the Presence of FrequencyOffset and a Method to Reduce Performane Degradation”Global Telecommunications Conference, 2000.GLOBECOM '00. IEEE,vol. 1,pp. 72-76, 2000.

[7] Yu Fu and Chi Chung Ko, “A New ICI Self –CancellationScheme for OFDM Systems based on a Generalized SignalMapper”, 5th International Symposium on Wireless PersonalMultimedia Communications, vol. 3, pp. 995-999, Oct. 2002.

[8] J. H. Mathews, K. K. Fink, Numerical Methods usingMatlab, 4th edition, Prentice-Hall Inc. , 2004.

6. CONCLUSIONSThe proposed scheme very well improves the CIR

performance of the OFDM system without increasinghardware complexity. The proposed sub optimalscheme completely removes the requirement of CFOestimation. However, the proposed scheme is slightlyless efficient than conventional SSR ICI selfcancellation in terms of BER.

(issn:2348-0246(online)),volume-1,issue-1,january ...ijetts.org/admin/issues/ijetts-010104.pdf ·...

Documents