Peak-to-Average Power ratio of IEEE 802.11a PHY layer Signals



Similar documents
Efficient Data Recovery scheme in PTS-Based OFDM systems with MATRIX Formulation

PAPR Reduction for 3FPP LTE OFDMA System

INTER CARRIER INTERFERENCE CANCELLATION IN HIGH SPEED OFDM SYSTEM Y. Naveena *1, K. Upendra Chowdary 2

Performance of Quasi-Constant Envelope Phase Modulation through Nonlinear Radio Channels

SC-FDMA and LTE Uplink Physical Layer Design

A Performance Study of Wireless Broadband Access (WiMAX)

HD Radio FM Transmission System Specifications Rev. F August 24, 2011

PAPR and Bandwidth Analysis of SISO-OFDM/WOFDM and MIMO OFDM/WOFDM (Wimax) for Multi-Path Fading Channels

Module: Digital Communications. Experiment 784. DSL Transmission. Institut für Nachrichtentechnik E-8 Technische Universität Hamburg-Harburg

Department of Electrical and Computer Engineering Ben-Gurion University of the Negev. LAB 1 - Introduction to USRP

Public Switched Telephone System

Optical Communications Research Group Department of Electronic and Computer Engineering University of Limerick, Ireland b

Time and Frequency Domain Equalization

Non-Data Aided Carrier Offset Compensation for SDR Implementation

802.11a White Paper. Table of Contents. VOCAL Technologies, Ltd. Home page

Digital Modulation. David Tipper. Department of Information Science and Telecommunications University of Pittsburgh. Typical Communication System

MODULATION Systems (part 1)

ISDB-T Digital terrestrial broadcasting in Japan

Adjacent Channel Interference. Adaptive Modulation and Coding. Advanced Mobile Phone System. Automatic Repeat Request. Additive White Gaussian Noise

Hot Issues in Wireless Broadband Networking

AN INTRODUCTION TO DIGITAL MODULATION

LTE Uplink Transmission Scheme

ADVANCED APPLICATIONS OF ELECTRICAL ENGINEERING

Performance of Multicast MISO-OFDM Systems

Vector Signal Analyzer FSQ-K70

An Investigation into Time-Domain Approach for OFDM Channel Estimation

Ralph L. Brooker, Member, IEEE. Andrew Corporation, Alexandria, VA 22314, USA

A PEAK TO AVERAGE POWER RATIO (PAPR) REDUCTION IN OFDM SYSTEMS

Evolution from Voiceband to Broadband Internet Access

Main Ways to Enhance Throughput

On the Influence of the Prediction Horizon in Dynamic Matrix Control

BROADCAST Trasmettitori / Transmitters DBH3 GAP FILLER

VDSL2 A feasible Solution for Last Mile

How To Encode Data From A Signal To A Signal (Wired) To A Bitcode (Wired Or Coaxial)

PHASE ESTIMATION ALGORITHM FOR FREQUENCY HOPPED BINARY PSK AND DPSK WAVEFORMS WITH SMALL NUMBER OF REFERENCE SYMBOLS

Introduction to Wireless LAN Measurements

Measurement, Modeling and Simulation of Power Line Channel for Indoor High-speed Data Communications

Noise Power and SNR Estimation for OFDM Based Wireless Communication Systems

Chap#5 (Data communication)

Plastic Optical Fiber for In-Home communication systems

CNR Requirements for DVB-T2 Fixed Reception Based on Field Trial Results

18-759: Wireless Networks Lecture 18: Cellular. Overview

4 Digital Video Signal According to ITU-BT.R.601 (CCIR 601) 43

Design of Bidirectional Coupling Circuit for Broadband Power-Line Communications

ISI Mitigation in Image Data for Wireless Wideband Communications Receivers using Adjustment of Estimated Flat Fading Errors

Two-slot Channel Estimation for Analog Network Coding Based on OFDM in a Frequency-selective Fading Channel

Suppression of Four Wave Mixing in 8 Channel DWDM System Using Hybrid Modulation Technique

Lezione 6 Communications Blockset

Throughput Modeling in IEEE WLAN-based Wireless Networks

Implementing Digital Wireless Systems. And an FCC update

Digital Subscriber Line (DSL) Transmission Methods

Study on Performance Analysis of CMOS RF front-end circuits for 2.4GHz Wireless Applications

Optimum Frequency-Domain Partial Response Encoding in OFDM System

INTRODUCTION TO COMMUNICATION SYSTEMS AND TRANSMISSION MEDIA

Chapter 28B - EMF and Terminal P.D. A PowerPoint Presentation by Paul E. Tippens, Professor of Physics Southern Polytechnic State University

The Effect of Network Cabling on Bit Error Rate Performance. By Paul Kish NORDX/CDT

Implementation of Digital Signal Processing: Some Background on GFSK Modulation

Mobile Communications Chapter 2: Wireless Transmission

Large-Capacity Optical Transmission Technologies Supporting the Optical Submarine Cable System

Dream DRM Receiver Documentation

Wireless Communication and RF System Design Using MATLAB and Simulink Giorgia Zucchelli Technical Marketing RF & Mixed-Signal

BENEFITS OF USING MULTIPLE PLP IN DVB-T2

On-Line Diagnosis using Orthogonal Multi-Tone Time Domain Reflectometry in a Lossy Cable

MIMO detector algorithms and their implementations for LTE/LTE-A

DVB-T BER MEASUREMENTS IN THE PRESENCE OF ADJACENT CHANNEL AND CO-CHANNEL ANALOGUE TELEVISION INTERFERENCE

Bachelor of Technology (Electronics and Communication Engineering)

Adaptive Techniques for Multiuser OFDM

Sampling Theorem Notes. Recall: That a time sampled signal is like taking a snap shot or picture of signal periodically.

RF Measurements Using a Modular Digitizer

Evaluation Criteria for ADSL AFE1302

What s The Difference Between Bit Rate And Baud Rate?

ISDB-T T Transmission Technologies and Emergency Warning System

RECOMMENDATION ITU-R F.1101 * Characteristics of digital fixed wireless systems below about 17 GHz

1 Multi-channel frequency division multiplex frequency modulation (FDM-FM) emissions

Achieving New Levels of Channel Density in Downstream Cable Transmitter Systems: RF DACs Deliver Smaller Size and Lower Power Consumption

5 Signal Design for Bandlimited Channels

DOCSIS 3.1: WHAT IS IT, HOW DOES IT WORK, AND HOW CAN YOU PREPARE FOR IT? Daniel Howard, SCTE CTO July 16, 2013

OFDM, Mobile Software Development Framework

AEROSOL STATISTICS LOGNORMAL DISTRIBUTIONS AND dn/dlogd p

SC-FDMA for 3GPP LTE uplink. Hong-Jik Kim, Ph. D.

MSB MODULATION DOUBLES CABLE TV CAPACITY Harold R. Walker and Bohdan Stryzak Pegasus Data Systems ( 5/12/06) pegasusdat@aol.com

Taking the Mystery out of the Infamous Formula, "SNR = 6.02N dB," and Why You Should Care. by Walt Kester

MATRIX TECHNICAL NOTES

TCOM 370 NOTES 99-4 BANDWIDTH, FREQUENCY RESPONSE, AND CAPACITY OF COMMUNICATION LINKS

Rapid Prototyping of a Frequency Hopping Ad Hoc Network System

An Algorithm for Automatic Base Station Placement in Cellular Network Deployment

Appendix D Digital Modulation and GMSK

Chapter 3: Spread Spectrum Technologies

MoCA 1.1 Specification for Device RF Characteristics

802.11ad - WLAN at 60 GHz A Technology Introduction White Paper

AN Application Note: FCC Regulations for ISM Band Devices: MHz. FCC Regulations for ISM Band Devices: MHz

Transcription:

Peak-to-Averae Power ratio of IEEE 802.a PHY layer Sinals A. D. S. Jayalath and C. Tellambura School of Computer Science and Software Enineerin Monash University, Clayton, VIC, 3800. e-mailjayalath@csse.monash.edu.au Abstract In this paper, we propose clippin with amplitude and phase chanes of the sinal points to reduce the peak-to-averae power ratio (PAR) of orthoonal frequency division multiplexed (OFDM) sinals in hih-speed wireless local area networks defined in IEEE 802.a physical layer. The proposed technique can be implemented with a small modification at the transmitter and the receiver remains standard compliant. PAR reduction as much as db can be achieved by selectin a suitable clippin ratio and a correction factor dependin on the constellation used. Out of band noise (OB) is also reduced. Key WordsOFDM, Peak-to-averae power ratio, IEEE 802.a PHY. I. ITRODUCTIO With a rapidly rowin demand for wireless communications, much research has been expended on providin efficient and reliable hih-data-rate wireless services. The IEEE 802. standard for wireless local area networks (WLA) was first established in 997, and it supported data rates of Mb/s and 2Mb/s in indoor wireless environments. In comparison to 00Mb/s Ethernet, the 2Mb/s data rate is relatively slow and is not sufficient for most multimedia applications. Recently, the IEEE 802. WLA standard roup finalized the IEEE Standard 802.a, which has an OFDM physical layer for indoor wireless data communications []. The data rates of IEEE 802.a rane from 6Mb/s up to 5Mb/s. This new standard can provide almost all multimedia communication services in indoor wireless environments. However, two limitations of OFDM based systems are often asserted. Firstly, due to the nonlinearities of transmitter power amplifiers, hih PAR values of the OFDM sinal enerates hih OB. There are restrictions imposed by the Federal Communications Commission and other reulatory bodies on the level of these spurious transmissions. These restrictions impose a maximum output power limitation. This output power limitation corresponds to what is known as power amplifier back-off. Further more nonlinearities of amplifier cause inband distortion of the sinal ivin hiher bit error rates (BER). 6-quadrature amplitude modulation (6-QAM) modulation is used in IEEE.a PHY for 8Mb/s and 5Mb/s transmission. This modulation scheme is hihly sensitive to distortions due small Euclidian distances between sinal points. Thus IEEE 802.a WLA devices may need power amplifiers with lare back off, which are inefficient and bulky. Overcomin this problem requires reducin the PAR of OFDM sinals. Secondly, OFDM is hihly sensitive to frequency off-set errors. However, frequency offset due to the mobility of the terminal is neliible in WLA environments, owin to very low speeds of the mobile. Clippin of hih peaks increases the in-band distortion and the OB. In [2] several modulation schemes are examined and effects of amplitude limitin is presented. A controlled amount of limitin is permissible in many cases. Clippin an oversampled sinal produces less in-band noise but the OB will increase [3]. Performance evaluation of clipped OFDM symbol with and without oversamplin is presented in []. Two extra oversampled DFT operation are used in the transmitter to filter the OB due to clippin. Oversamplin is necessary to avoid peak rerowth after filterin. Similar PAR reduction scheme based on clippin and filterin an oversampled sinal is presented in [5]. Two oversampled inverse fast Fourier transform (IFFT) operations are used to interpolate and perform frequency domain filterin after clippin. This reduces the amount of OB. Increased BER due to clippin may be reduced usin forward error correction codin. Thus, a system based on clippin and forward error correction is proposed in [6]. Decision aided reconstruction (DAR) is proposed in [7] for mitiatin the clippin noise. The receiver is assumed to know the clippin level. DAR is an iterative reconstruction technique, which increases the receiver complexity. Several discrete Fourier transform (DFT) operations are performed at the receiver before reconstructin the oriinal sinal. For baseband transmission, an approximation for the resultin increase in BER is iven in [8], but spectral distribution of the distortion is not considered. References [9, 0] consider the spectral spreadin but only for real valued discrete multitone (DMT) sinals. The deradation due to clippin of an OFDM sinal is analyzed in []. Extensive research has been undertaken to understand the effect of hih power amplifier (HPA)

G I nonlinear distortion in OFDM systems [2, 3]. Compared to other PAR reduction schemes clippin based techniques are simple to implement. A PAR reduction usin sinal clippin and phase correction at the transmitter is proposed in []. The basic idea is as follows. Let the IFFT of the modu- lated symbol sequence be. Here we assume quadrature phase shift keyin (QPSK). The sequence is clipped to a desired level, and the IDFT is taken. Because "! # of the clippin. The key idea in [] is to apply phase correction to such that both and have identical phases. This is done by multiplyin with $%&('*),+.- '*)0/ where and are the phases of and. The idea of this phase correction is to limit OB and BER deradation. Of course, we cannot make the amplitude equal to the oriinal sinal manitude. Then we would be back to the oriinal input sequence. Finally, the IDFT of the phase-corrected sequence is used to enerate the transmitted sinal. ote that only samples per each OFDM symbols are enerated. Clippin those to a low level does not uarantee that the peaks of the output sinals are reduced to the same level. This paper presents three enhancements to the method proposed in []. First, we oversample the OFDM sinal by a factor of 2. Oversamplin will lead to reater robustness aainst peak rerowth. Oversamplin is implemented usin 2 -lenth IDFTs. Secondly, the phase correction is controlled (additional parameter 3 ) and is applied only if the difference between and exceeds 3. Third, if the amplitude deviates too much from the oriinal sinal amplitude (as measure by a parameter ), an amplitude adjustment is applied. The BER deradation caused is thus neliible and the OB is reduced considerably. This approach is more eneral than []. For instance the special case in our scheme represented by 65 and 3 =0 corresponds to []. We then apply this new scheme in IEEE 802.a PHY transmit sinals and evaluate the performance. II. A OFDM SYSTEM AD PEAK-TO-AVERAGE POWER RATIO (PAR) 7 8 "9 A block of symbols,, is 7 ; 9 formed with each symbol modulatin one of a set of subcarriers with frequency,. The subcarriers are chosen to be orthoonal, that is 8 =< 8?>A@ @, where is the OFDM symbol duration. The complex baseband sinal can be expressed as?bdc E F +H KJML $ %OP*)0Q SR C RT@U () FEC CODER Interleavin + Mappin Oversampled IFFT L ew block AMP and Phase chane A -A Soft limiter (SL) L IFFT Oversampled FFT Symbol Wave Shapin Amplitude and Phase Adjustments Fi.. IEEE 802.a Transmitter. This sinal can be enerated by takin an point V inverse discrete Fourier transform (IDFT) of the block followed by low pass filterin. The actual transmitted sinal is modelled as real W?BDC E $%OP X*QZY, where K[ is the carrier frequency. The PAR of the transmitted sinal in () can be defined as \ ^] c d _0`ba?BDC E a (2) a?bdc E a e c d where e is the expected value of. The PAR of the continuous-time OFDM sinal cannot be computed precisely by the use of the yquist samplin rate [5], which amounts to samples per symbol. In this case, sinal peaks are missed and PAR reduction estimates are unduly optimistic. Oversamplin by a factor of is sufficiently accurate. III. IEEE 802.f SYSTEM DESCRIPTIO The IEEE 802. transmitter with the proposed PAR reduction scheme is presented in Fiure. Input data is first mapped into symbols ( ) and serial to parallel converted. Then a 2 point oversampled IFFT is taken. The output of oversampled IFFT is clipped accordin the clippin ratio selected. A second 2 point fast Fourier transform (FFT) is taken to et the clipped sample points back to frequency domain. We will now select samples correspondin to the oriinal sinals ( ). The clippin of the sinal causes dispersion of the sinal points. If we do not clip the sinal, the oriinal symbols ( ) will be reenerated at this point. ext, we adjust the phase and the amplitude of these sinal points such that they are confined V to a smaller reion around the oriinal sinal points. Let us define an arbitrary sinal point of the iven constellation havin amplitude and phase and as shown in Fiure 2. Then phase, and amplitude of each sinal point is adjusted accordin to the followin rules. If the amplitude of deviates from by more than, the corrected amplitude will be adjusted as follows. ihj k lj bm n n p n else.o.o IQ Mod (3)

3 lj 3 3 3 L Fi. 2. Amplitude and phase adjustment parameters. where. If the phase of the clipped sinal deviates by 3 or more from the phase of the oriinal sinal points, the phase is also adjusted. 3 is calculated from the iven as follows k hj where _ +H m Vo 3 Vo else () (5). At this point we have a new symbol sequence optimized for lower PAR. This new scheme needs two extra oversampled-ifft operations at the transmitter. However, o modifications are needed at the receiver. TABLE I IEEE 802. PHY PARAMETERS Information data rate 6,9,2,2,36 8 and 5 Mb/s Modulation BPSK-OFDM QPSK-OFDM 6-QAM-OFDM 6-QAM-OFDM FEC code Convolutional rate /2, (K=7) Code rates /2, 2/3,3/ Total umber of sub carriers 52 umber of pilot subcarriers OFDM symbol duration Guard interval 0.8 Sinal bandwidth 6.6 MHz IV. CLIPPIG HIGH PEAKS A soft limiter (SL) described below is used to clip the sinal peaks. The nonlinear characteristics of an SL can Fi. 3. Sinal distortion at db clippin of 8Mb/s Sinals. be written as where $% a a a a R o SR VR ^ (6) is the clippin level and is the phase anle of the sample M and is the clipped output sequence. The clippin ratio is defined as Clippin ratio H V. RESULTS c d a a e db (7) Simulation results were obtained for different transmit sinals. As IEEE 802.a transmit sinals utilize four constellations (binary phase shift keyin (BPSK), QPSK, 6-QAM and 6-QAM), simulations are performed for data rates involvin these constellations except for BPSK. Clippin has little effect on BPSK OFDM. Fiure 3 depicts the constellation of a 8Mb/s transmitted sinal passin throuh a SL with a db back off. The distortion of the sinal is clearly evident. Fiure shows the amplitude and phase corrected sinal after db clippin with correction factor ;. The distortion of the sinal reduces sinificantly dependin on the value of the correction factor. Fiure 5 depicts the complementary cumulative distribution function of the PAR of the 8Mb/s sinal. PAR is reduced by more than db with the db clippin, but with a severely distorted sinal. Phase and amplitude correction reduces the distortion of the sinal points but derades the PAR statistics. A 3dB ain in PAR statistics is obtained when, while this ain is around 2dB when at + of CCDF. Clippin with suitable phase and amplitude correction can achieve a sinificant PAR reduction. We also examined the performance of other schemes. Fiure 6 depicts the 36Mb/s sinal constellation after passin throuh a SL with 2dB clippin. As 6-QAM

Fi.. Corrected sinal constellation at db clippin (8Mb/s and ). Fi. 6. Sinal distortion at 2dB clippin of 36Mb/s Sinals. 0 0 0 Without adjustments Without clippin 0 2 alpha=0.02 fra replacements alpha=0.05 0 3 2 3 5 6 7 8 9 0 2 Fi. 5. CCDF of the 8Mb/s sinals with db clippin. is less condensed than the constellation for 8Mb/s or 5Mb/s (6-QAM), we can allow for hiher clippin ratios. Fiure 7 shows the amplitude and phase corrected sinal constellation. The amplitude and phase distortion is reatly reduced when the correction factor. Fiure 8 depicts the CCDF of PAR of the transmitted sinals. The PAR statistics improve as before. The correction factor is chosen as and with clippin at 2dB. These parameters correspond to PAR statistics improvements of 3dB and db respectively. When the simulations are performed for 2Mb/s sinals we could observe far better performance improve- ments. The constellation in 2Mb/s sinal is less dense than the two other schemes before and allows reater flexibility in selectin correction factor PSfra. Hih replacements values of ive out hih PAR reduction. When the ; ^ about 5dB ain is observed while this was more than 3dB when. ext we observe the power spectral density (PSD) of the transmitted sinal when passin throuh a non- Fi. 7. Corrected sinal constellation at 2dB clippin (36Mb/s and ). 0 0 0 0 2 Without adjustments alpha=0.5 alpha=0. Without clippin 0 3 0 2 6 8 0 2 Fi. 8. CCDF of the 36Mb/s sinals with 2dB clippin.

@ 5 e e 0 0 Ordinary Sinals Modified Sinals 20 30 3dB Power spectral density 0 50 60 70 7dB 7dB 5dB 5dB 9dB 3dB 80 90 9dB 00 3 2 0 2 3 ormalized frequency B n Fi. 9. Sinal distortion at 2dB clippin of 2 Mb/s Sinals. Fi. 2. Power spectral density when passin throuh a non-linear amplifier. fra replacements Fi. 0. Corrected sinal constellation at 2dB clippin (2Mb/s and ). 0 0 0 0 2 Without adjustments alpha=0.2 alpha=0. Without clippin 0 3 0 2 6 8 0 2 Fi.. CCDF of the 2 Mb/s sinals with 2dB clippin. linear power amplifier. The input output relationship of power amplifiers can be modelled as d d d m B E e 8 (8) where is the clippin level, is the amplitude of the sinal input and and denote the amplitude and the phase at the output respectively. The parameter controls the smoothness of the transition from the linear reion to the limitin or saturation reion. When, this model approximates the SL characteristics. For PSD results, it is convenient to define the normalized bandwidth B < > E @U>, where is the OFDM symbol duration. In Fiure 2, 2 Mb/s sinal with 2dB clippin and correction factor is compared with an ordinary transmitted sinal. The back-off of the power amplifier is set at different levels (3dB, 5dB, 7dB and 9dB). When the back off is very low (3dB) very sliht improvement in OB is observed. OB is reduced by about 8dB when the back-off is at 5dB. Similarly, OB reduces sinificantly with the increase in back-off at a faster rate compared to the ordinary sinal. Therefore the back-off of the amplifier can be reduced sinificantly by usin this technique. Selection of initial clippin ratio and the correction factor depends on the sinal constellation bein used. Therefore, by selectin a suitable clippin level and a proper phase an amplitude correction factor sinificant PAR reduction is achieved without causin sinificant BER deradation. This is a desired feature in portable devices in WLAs, where power efficient transmitter power amplifiers are essential.

VI. COCLUSIO A technique based on clippin with amplitude and phase chanes to reduce the PAR of OFDM based WLA sinals defined in the IEEE 802.a physical layer is presented in this paper. The proposed technique is capable of reducin the PAR by 3-dB by selectin a suitable clippin level and amplitude and phase correction factor. It can be implemented with a sliht increase in the complexity at the transmitter. This involves with insertion of two additional DFT operations and soft limiter at the transmitter. The receiver remains standard compliant. REFERECES [] IEEE Std 802.a-999 (supplement to IEEE Std 802.-999, Part II Wireless LA medium access control (MAC) and physical layer PHY specifications Hih speed physical layer in 5 GHz bnad, tech. rep., IEEE, Sept. 999. [2] D. W. Bennett, P. B. Keninton, and R. J. Wilkinson, Distortion effects of multicarrier envelope limitin, IEE Proc. Commun., vol., pp. 39 356, Oct. 997. [3] X. Li and L. J. Cimini, Effects of clippin and filterin on the performance of OFDM, IEEE Commun. Lett., vol. 2, pp. 3 33, May 998. [] H. Ochiai and H. Imai, On the clippin for peak power reduction of OFDM sinals, in IEEE GLOBECOM, (San Fracisco, USA), pp. 73 735, IEEE, 2000. [5] J. Armstron, ew OFDM peak-to-averae power reduction scheme, in IEEE Vehicular Technoloy Conference, (http//www.ee.latrobe.edu.au/ ja/pubs.htm), IEEE, May 200. [6] D. Wulich and L. Goldfeld, Reduction of peak factor in orthoonal multicarrier modulation by amplitude limittin and codin, IEEE Trans. Commun., vol. 7, pp. 8 2, Jan. 999. [7] D. Kim and G. L. Stuber, Clippin noise mitiation for OFDM by decision-aided reconstruction, IEEE Commun. Lett., vol. 3, pp. 6, Jan. 999. [8] R. O eill and L. B. Lopes, Performance of amplitude limited multitone sinals, in IEEE Vehicular Technoloy Conference, (ew York, Y, USA), pp. 675 679, IEEE, 99. [9] D. J. G. Mestdah, P. Spruyt, and B. U. Brain, Analysis of clippin effect in DMT-based ADSL systems, in IEEE ICC, (ew York, Y, USA), pp. 293 300, IEEE, 99. [0] R. Gross and D. Veeneman, SR and spectral properties for a clipped DMT ADSL sinal, in IEEE ICC, (ew York, Y, USA), pp. 83 87, IEEE, 99. [] M. Friese, On the deradation of OFDM-sinals due to peakclippin in optimally predistorted power amplifiers, in IEEE GLOBECOM, (Sydney, Australia), pp. 936 9, IEEE, 998. [2] E. Boenfeld, R. Valentin, K. Metzer, and W. Sauer-Greff, Influence of nonlinear HPA on trellsi-coded OFDM for terrestrial broadcastin of diital HDTV, in IEEE GLOBECOM, (ew York, Y, USA), pp. 33 38, IEEE, 993. [3] G. Santella and F. Mazzena, A hybrid analytical-simulation procedure for performance evaluation in M-QAM-OFDM schemes in presence of nonlinear distortions, IEEE Trans. Veh. Technol., vol. 7, pp. 2 5, Feb. 998. [] M. C. D. Maddocks and P. Shelswell, Shelswell Diital sinal transmission system usin frequency division multiplex. U. S. Patent U. S. Patent US 5 60 908, 997. [5] C. Tellambura, Phase optimization criterion for reducin peakto-averae power ratio in OFDM, IEE Elect. Lett., vol. 3, pp. 69 70, Jan. 998.