The goal for the future wireless communications system

The goal for the future wireless communications system is to seamlessly integrate awide variety of communication services.The current focus is to design systems that supportwideband services like high speed Internet access, video and high-quality image transmissionwith the same quality as the fixed networks. Orthogonal Frequency Division Multiplexing(OFDM) is part of the wireless LAN standards IEEE 802.11a, g, n and y, DAB-T, and DVBT.In addition, OFDM is specified in the High-Performance LAN (HIPERLAN/2), in theFixed Worldwide Interoperability for Microwave Access (WiMAX), which is based, on theIEEE 802.16-2004 standard and the mobile WiMAX, which is based on the IEEE 802.16e-2005 standard.OFDM is a multicarrier technique in which a high-bit-rate data stream is divided into severalparallel lower bit-rate streams by modulating each stream on separate carriers calledsubcarriers as shown in Fig.1.1.The performance of the OFDM system is also critically dependent on the orthogonalityof the subcarriers. Therefore, frequency shifts and timing errors result in substantial Bit ErrorRate (BER) degradation. Synchronization between transmitter and receiver is crucial.Frequency shifts at transmitting or receiving end can lead to the displacement of OFDMsignal in spectrum and/or inter carrier interference (ICI).Furthermore, the structure of the OFDM signal gives rise to a possibility that all subcarrierspower adding up constructively leading to a combined signal with a very high peak power andhence a high Peak to Average Power Ratio (PAPR). A high PAPR places a constraint on thedesign of the transmitter as it dictates the dynamic range of its circuitry. In case the circuitryis unable to cater for the high PAPR, the resultant signal becomes distorted.2Fig.1.1 Block diagram of a OFDM transmitter and receiver system1.1.1 Advantages1. Makes efficient use of the spectrum by allowing overlap.2. By dividing the channel into narrowband flat fading subchannels, OFDM is moreresistant to frequency selective fading than single carrier systems are.3. Eliminates ISI and IFI through use of a cyclic prefix.4. Using adequate channel coding and interleaving one can recover symbols lost due tothe frequency selectivity of the channel5. Channel equalization becomes simpler than by using adaptive equalization techniqueswith single carrier systems.6. It is possible to use maximum likelihood decoding with reasonable complexity.7. OFDM is computationally efficient by using FFT techniques to implement themodulation and demodulation functions.8. Is less sensitive to sample timing offsets than single carrier systems are.9. Provides good protection against cochannel interference and impulsive parasitic noise.31.1.2 Disadvantages1. The OFDM signal has a noise like amplitude with a very large dynamic range,therefore it requires RF power amplifiers with a high peak to average power ratio.2. It is more sensitive to carrier frequency offset and drift than single carrier systems aredue to leakage of the DFT.PAPR reduction methods have been studied for many years 2 and significant number ofmethods has been developed. A PAPR reduction technique in which side information is notneeded to be transmitted is said to be bandwidth efficient. A PAPR reduction technique inwhich either Nonlinear distortion and processing time & power consumption of processor orboth is reduced is said to be power efficient.• Clipping: Clipping naturally happens in transmitter if power back-off is not enough.Clipping leads to a clipping noise and out-of-band radiation. Filtering after clipping canreduce out-of-band radiation, but at the same time it can cause “peak regrowth”. Repeatedclipping and filtering can be applied to reduce peak regrowth in expense of complexity.Several methods for mitigation of the clipping noise at receiver were proposed.• Coding: Coding methods include Golay complementary sequences, simple block codingscheme, complementary block codes (CBC), modified complementary block codes(MCBC) etc. An application of the Coding methods have low complexity but PAPRreduction is achieved in expense of redundancy causing data rate loss.• Partial Transmit Sequences (PTS): A set of sub-carriers of an OFDM symbol is dividedinto V non-overlapping sub-blocks. Each sub-block undergoes zero-padding by insertingzeroes at those sub-carriers which are already represented in other sub-blocks. All V subblocksare transformed into the time-domain by IDFT resulting in V partial transmitsequences denoted by p(k), k=1…V. The transmit sequence is obtained as a linearcombination of PTSs:??Vkp k b k1( ) ( ) , where b(k) is a rotation factor.The optimization isperformed over the rotation factors b(k) for each OFDM symbol. It is shown that fourrotation angles (b(k)???1? j?) is already enough for significant peak power reduction.6• Selected mapping (SLM): This method defines U distinct rotation vectors. The length ofthese vectors is equal to the number of sub-carriers of the OFDM symbol. The OFDMsymbol is multiplied sub-carrier wise by rotation vector U1 and transformed into the timedomainby IDFT. Then the same procedure is repeated for vectors U2, U3…. At the endthe vector with the lowest PAPR is selected for transmission.• Interleaving: The same data block is interleaved by K different interleavers. K-IDFTs ofthe original data block and modified data blocks are calculated. PAPR of K blocks iscalculated. The block with lowest PAPR is transmitted.• Tone Reservation (TR): L sub-carriers are reserved for peak reduction purposes. Thevalues of the signals to insert on peak reduction sub-carriers are computed by suitableLinear Programming algorithm.• Tone Injection (TI): TI maps one constellation point of the original constellation (forexample QPSK) to several constellation points of the expanded constellation (for example16QAM). PAPR reduction is achieved by choosing constellation points of the expandedconstellation.• Active Constellation Extension (ACE): ACE modifies original constellation by movingnominal constellation points located on the outer constellation boundaries in the directionsthat don’t decrease Euclidean distances between constellation points.• Nonlinear Companding Transform (NCT): NCT companded original OFDM signal usingstrict monotone increasing function, which enlarges the small signals and compresses thelarge signals. Companded signal can be recovered by the inverse function at receiver.One of the major issues in an OFDM signal generated at the transmitter is that there is alikelihood that all subcarriers present at the IFFT input may be in phase. This may result inthe subcarrier powers adding up, constructively leading to a combined signal with very highpeak power resulting in high PAPR. SLM is an alternative solution proposed to reduce PAPRbut the PAPR reduction is based on proper choice of phase sequence sets.8Secondly, in SLM, orthogonal or nonorthogonal phase sequences are used to generatecandidate signal copies. An orthogonal set does not need the transmission of side information,thereby ensuring bandwidth efficiency whereas it is not so for nonorthogonal sequences.Component wise multiplication of input sequence with orthogonal and aperiodic phasesequences results in PAPR reduction.The other major issue of concern is when such high PAPR signal is transmitted through theHigh-Power Amplifier (HPA), it is forced to operate with large power back-off to avoid nonlineardistortions. Despite PAPR reduction the third order Inter-modulation Distortion(IMD3)products causes out of band distortion and in band distortion. Odd-order products introducein-band distortion and even-order products introduce out-of-band distortion. If in banddistortion is not reduced in SLM systems, it causes BER performance degradation, whereasout of band distortion results in spectral regrowth.Error vector magnitude (EVM) is widely adopted to quantify the amount of in-band distortionthat occurs at OFDM transmitters and the constellation error exhibited on the received signal.It directly affects the error performance, e.g., bit error rate, thereby reducing power efficiencyof the transmitter.Spectral regrowth is IMD3 energy leak to the adjacent channels and in measured by ACPR(Adjacent Channel Power Ratio).Another major drawback with SLM systems is the high computational complexity. Manymodified SLM schemes with low complexity have been proposed 12 – 15, but they haveseveral shortcomings such as degradation of PAPR reduction performance or BERdegradation compared to the conventional SLM scheme using the same number of candidatesignal sequences.This is because in conventional SLM for each candidate signal mapping needs U*(1+N/2)log2(N)) complex multiplications and U*(N)log2(N) complex additions for IFFT operation.Further fixed number of U mappings are performed irrespective of whether PAPR is alreadyreaches below a certain set threshold in first few mappings or not. Hence complexityincreases as U increases.9Therefore, high PAPR in SLM systems places a constraint on the design of the transmitter asit decides the bandwidth efficiency, power efficiency of the HPA, the overall performance ofthe OFDM system. The current implementations of SLM-OFDM do not fully exploit itscapabilities. There are still several avenues that can be explored to reduce the peak-to-powerratio (PAPR) of the OFDM signal. Therefore, the necessity to propose a bandwidth efficient,power efficient, reduced PAPR SLM-OFDM technique is a prime motivation for the researchwork.1.4 ApplicabilityThe air interface requirements for PAPR reduction to be applicable for communicationnetworks such as 802.11a, 802.11n, 802.16m, 802.11p, or Evolved Universal TerrestrialRadio Access Network (E-UTRAN) should satisfy several requirements:1. No/minimum data rate loss.2. No spectrum spillage.3. No Average power increase.4. No BER performance degradation.5. Moderate complexity.1.5 Problem DefinitionOFDM is computationally efficient by using FFT techniques to implement the modulationand demodulation functions but the time domain OFDM signal has a noise like amplitudewith a very large dynamic range i.e. OFDM signals suffer from large envelope variations.Practical systems being peak power limited, such large envelope variations are problematicand the issue needs to be addressed. Thus, envelope peaks require a modification to existingOFDM system to accommodate an instantaneous signal power that is larger than the signalaverage power, necessitating low PAPR systems to ensure linear operation of RF amplifierswithout degradation in power efficiency. We intend to explore the existing PAPR reductiontechniques and to find out their advantages and major limitations for implementing a practicalOFDM system. SLM and PTS are two popular PAPR reduction techniques but PTS hashigher computational complexity than SLM. The aim is to propose an efficient PAPR10reduction applicable to wireless systems such as WLAN, WiMAX or LTE for single andmultiple antenna systems.In this thesis we investigate and propose modifications to existing SLM and to achievebesides PAPR reduction provides an improvement in one or more of the following to make itPower efficient.1. Reduced Nonlinear distortion2. Reduced Number of complex additions and multiplications required perOFDM frame1.6 Objectives1. The prime objective is to design new phase sequence set that is crucial to thepercentage of PAPR reduction achievable in SLM for SISO and STBC -MIMOOFDM systems.2. Improve the correlation properties of the entire set and of individual sequencesapplying parameters such asa. Aperiodic auto and cross correlationb. Merit factorc. Average mean-square value of auto correlationd. Maximum value of aperiodic cross correlation function.3. To reduce Nonlinear distortion to improve power efficiency for the proposed SLM bysubblock partitioning and precoding. Further, we analyze quantitatively effect ofreduction in nonlinear distortion on BER.4. To reduce complexity and save processing time & power consumption of processor byapplying serial approach based SLM for orthogonal phase sequences. Further, toanalyze the significant decrease in the average number of assessed candidate signalsper OFDM frame.In chapter 2, we identified and studied the problem that arises from high PAPR by review ofvarious existing literature.Chapter 3, first will introduce the details of SLM and phase sequence sets. Next, weintroduce a new phase sequence set and discuss its correlation properties to decide on itssuitability to SLM. Finally, we will provide a detailed discussion on results based on proposedphase sequence set. The performance of the proposed system is analyzed in terms of PAPR,BER, EVM and ACPR for SISO and MIMO systems.Chapter 4 gives an overview of techniques such as precoding and subblock partitioning toreduce in band distortion which further reduces the SNR requirement of SLM systems toovercome BER degradation and achieve improved power efficiency.In chapter 5, serial approach to SLM referred to as Iterative SLM for improved powerefficiency is discussed in detail. Conventional SLM for each candidate signal mapping needsU*(1+N/2) *log2(N)) complex multiplications and U*(N)log2(N) complex additions forIFFT operation. Further U mappings are performed irrespective of whether PAPR may12already be reduced below a certain threshold in first few mappings or not. Hence complexityincreases as U increases increasing power consumption of the processor.Chapter 6 discusses the various results obtained and validated with literature whileChapter 7 lists the author’s contribution to research and concludes the report.