Unveiling Myths about SC-FDMA in TGm IEEE 802.16 Presentation Submission Template (Rev. 9) Document Number: IEEE C802.16m-08/066 Date Submitted: 2008-01-16 Source: Thierry Lestable, Alain Mourad, Ming Jiang, Youngkwon Cho Voice: +44. 1784.428600 Ext 720 Samsung Electronics Research Institute, UK E-mail: {thierry.lestable, alain.mourad, ming.jiang, youngkn}@samsung.com Junsung Lim, Hokyu Choi Samsung Electronics Co., Ltd. 416 Maetan-3, Suwon, 442-600, Korea Voice: E-mail: {junsung.lim, choihk}@samsung.com Venue: IEEE 802.16m-07/047, Call for Contributions on Project 802.16m System Description Document (SDD). Base Contribution: None Purpose: To be discussed and adopted by TGm for the 802.16m SDD Notice: This document does not represent the agreed views of the IEEE 802.16 Working Group or any of its subgroups. It represents only the views of the participants listed in the Source(s) field above. It is offered as a basis for discussion. It is not binding on the contributor(s), who reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution, and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE s name any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE s sole discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The contributor also acknowledges and accepts that this contribution may be made public by IEEE 802.16. Patent Policy: The contributor is familiar with the IEEE-SA Patent Policy and Procedures: <http://standards.ieee.org/guides/bylaws/sect6-7.html#6> and <http://standards.ieee.org/guides/opman/sect6.html#6.3>. Further information is located at <http://standards.ieee.org/board/pat/pat-material.html> and <http://standards.ieee.org/board/pat >.
Unveiling Myths about SC-FDMA in TGm Thierry Lestable, Alain Mourad, Junsung Lim, Ming Jiang, Youngkwon Cho, Hokyu Choi Samsung Electronics Co., Ltd. January, 2008
Outline Introduction of SC-FDMA Key Considerations for TGm Complexity Out of Band Emissions Link Level Performance Multiplexing Pilot Tone Insertion System Level Preliminary Evaluations Conclusions & Proposal
Introduction of SC-FDMA
SC-FDMA Transmitter Properties of SC-FDMA Tx Single carrier transmission due to DFT spreading Lower PAPR at the cost of out of band emission Need pulse shaping filter : Back to high PAPR { } n x { } DFT (M-point) X k { l } Sub-carrier Mapping X { x } IDFT (L-point) m Add CP /pulse shaping { x n } x0 x1 x2 x3 T b T s I-FDMA { x } m x0 x1 x2 x3 x0 x1 x2 x3 x0 x 1 x2 x3 x0 x 1 x2 x3 L-FDMA { x } m x??? 0 x? 1?? x 2??? x? 3?? An example of SC-FDMA transmit symbols in the time domain for M=4, Q=4, and L=16.
SC-FDMA Transmitter - PAPR Pulse shaping incurs higher PAPR Comparison of CCDF of PAPR for IFDMA and LFDMA with M = 256, N = 64, and roll off factor(α) of 0, 0.2, 0.4, and 0.6, 0.8, and 1. (a) QPSK. (b) 16-QAM.
SC-FDMA Receiver Properties of SC-FDMA Rx Vulnerable to severe frequency selective fading Lower post-sinr than OFDMA 1-tap frequency domain equalization per subcarrier A block of input symbols experiences same distortion. One severe FDE loss is detrimental to a block information Should be taken into account for cell edge user { } y i DFT (L-point) { } l Y Y m Frequency { Z } { } De-map (M-point) { } Domain Equalizer k IDFT (M-point) z n
Key Considerations for TGm
Complexity Transmitter (MS side) Require additional DFT process with dynamic DFT size relying on number of allocated subcarriers. Receiver (BS side) Additional IDFT processing High complexity frequency domain equalizer Impractical implementation for ML receiver Consequently, Increase burden for detection and decoding. Additional power consumption Additional processing delay is prohibitive to handset.
Out of Band (OOB) Emissions SC-FDMA: Higher instantaneous out of band emission Interfere to adjacent channels OOB can be compensated with pulse shaping at the cost of PAPR. Waste of resource increasing guard band 10 0-10 -20-30 -40-50 Inst. PSD (4 symbols), N=1024, M=128 SC-FDMA OFDMA -60-2000 -1500-1000 -500 0 500 1000 1500 2000 subcarrier *Intel Corp. Ref. Alamouti presentation
Link Level Performance OFDMA outperforms SC-FDMA. OFDMA can have up to 5dB gain w.r.t. SC-FDMA for SIMO. The benefit of OFDMA becomes significant for MIMO. 2dB 6.5dB J. Zhang, et al., Comparison of the Link Level Performance between OFDMA and SC-FDMA, IEEE 2006
Required DFT Sizes SC requires variable DFT sizes depending on radio resources allocated LTE: radix-2, radix-3 and radix-5 n m p N DFT = 2 3 5 c.f.) 802.16e: radix-2, radix-3 1 subchannel / Nfft =2048 N DFT = n 48 = 2 3 LTE DFT DFT =36 DFT =24 =12 DFT =12 DFT =36 DFT =12 DFT =36 DFT =12 DFT =36
Insertion of Reference Signals Lack of flexibility Frequency domain insertion Back to high PAPR Time domain insertion Due to constraints on time multiplexing for reference signal, we end up with increased problems w.r.t. backward compatibility of frame structure. Less flexible on pilot arrangement. Channel estimation error becomes critical for SC-FDMA. Pilot pattern is likely to be even more sensitive to support collaborative SM. 9 DFT-spread data... data DFT-spreading Pilot tones Example of frequency domain insertion... f 99.9% PAPR [db] 8 7 6 5 4 3 2 1 0 OFDMA DFT-s- OFDMA_D (w/o pilot) DFT-s- OFDMA_L (w/o pilot) DFT-s- OFDMA_S (w/o pilot) DFT-s- OFDMA_D (w/ pilot) DFT-s- OFDMA_L (w/ pilot) QPSK 16QAM DFT-s- OFDMA_S (w/ pilot)
System Level Preliminary Evaluations
OFDMA Parameters Parameter Description Value [802.16m] F c Carrier frequency 2.5 GHz BW Total bandwidth 10 MHz N FFT Number of points in full FFT 1024 F s Sampling frequency 11.2 MHz Δf Sub-carrier spacing 10.9375 khz T 0 =1/ Δf CP N usc N scch OFDM symbol duration without cyclic prefix Cyclic prefix length (fraction of T 0 ) Number of used data subcarriers Number of used data subcarriers per sub-channel 91.43 μs N maxch Number of sub-channels 35 1/8 840 24
Test Scenarios Scenario / Parameter Baseline NGMN Urban Macrocell Requirements Mandatory [802.16m] Optional [802.16m] Optional [802.16m] Site-to-Site distance 1.5 Km 0.5 Km 1 Km Carrier frequency 2.5 GHz 2.5 GHz 2.5 GHz Operating Bandwidth 10 MHz 10 MHz 10 MHz MS Tx Power 23 dbm 23 dbm 23 dbm Penetration loss 10 db 20 db 10 db Path loss model Lognormal shadowing standard deviation Inter-site shadowing correlation PL (db) = 130.62 + 37.6log10(R) (R in km) PL (db) = 130.62 + 37.6log10(R) (R in km) PL (db) = 35.2 + 35log10(R) + 26log10(f/2) (R in meter, f in GHz) 8 db 8 db 8 db 0.5 0.5 0.5 Channel Mix ITU Veh A (30 km/hr) 100 % ITU Ped B (3 km/hr) 100 % ITU Veh A (30 km/hr) 100 %
System Parameters Parameter Value Number of sites 19 Number of sectors per site 3 Wrap-around technique Yes Frequency reuse 1 Number of MS Tx antennas 1 Number of BS Rx antennas 1 BS antenna pattern BS antenna gain MS antenna pattern MS antenna gain BS noise figure Thermal noise density 2 θ min 12, Am ; Am = 20 db, θ3 3 θ db 17 dbi Omi-directional 0 dbi 5 db -174 dbm/hz Number of sub-channels requested by each MS 1 Average number of MS per sector 5 Sub-carriers mapping Localized Receiver structure MMSE db = 70
CDF of SINR: NGMN Scenario 1 0.9 OFDMA (OBO = 0 db) OFDMA (OBO = 3 db) OFDMA (OBO = 6 db) SC-FDMA (MMSE) 1 0.9 0.8 OFDMA (OBO = 0 db) OFDMA (OBO = 3 db) OFDMA (OBO = 6 db) SC-FDMA (MMSE) 0.8 0.7 5 subchannels used per sector 0.6 0.7 CDF 0.5 Load = 5/35 = 15% 0.4 0.6 0.3 0.2 CDF 0.5 0.1 0.4 0-10 -5 0 5 10 15 20 25 30 35 40 IoT (db) 0.3 0.2 0.1 This is an interference limited scenario. OBO has little impact on OFDMA. OFDMA has close performance to SC-FDMA even for OBO = 6 db 0-20 -10 0 10 20 30 40 Effective SINR (db) Still Need to add Channel Estimation Errors Modelling This context is in favour of OFDMA
CDF of SINR: Baseline Scenario 1 0.9 OFDMA (OBO = 0 db) OFDMA (OBO = 3 db) OFDMA (OBO = 6 db) SC-FDMA (MMSE) 1 0.9 0.8 OFDMA (OBO = 0 db) OFDMA (OBO = 3 db) OFDMA (OBO = 6 db) SC-FDMA (MMSE) 0.8 0.7 20 subchannels used per sector 0.6 0.7 0.6 Load = 20/35 = 60% CDF 0.5 0.4 0.3 CDF 0.5 0.2 0.4 0.1 0.3 0 0 5 10 15 20 25 30 35 40 IoT (db) 0.2 0.1 0-30 -20-10 0 10 20 30 Effective SINR (db) Still Need to add Channel Estimation Errors Modelling This is interference limited scenario with 20 users/sector. OBO has little impact on OFDMA. OFDMA has close performance to SC- FDMA even for OBO = 6 db This context is in favour of OFDMA
Observations In interference-limited scenarios, OFDMA always achieves higher SINR values OBO does not degrade the performance of OFDMA in interference-limited scenarios For 5 users per sector (15% Load), NGMN is already interference limited For 20 users per sector (60% Load), baseline scenario is interference limited OFDMA is better suited to scenarios with low site-to-site distance and high sector loading/throughput
Summary & Conclusion Drawbacks of SC-FDMA Degrade link performance especially with high order modulation and MIMO in real channel estimation scenario. Additional out of band emission problematic with adjacent channel or bandwidth efficiency Additional complexity and power consumption on transmitter and receiver Lack of flexibility on pilot/reference signal insertion Conclusion & Proposal Sticking with dominant OFDMA basis is better to ensure: Easier & safer backward compatibility Reaching increased spectral efficiency Easy and flexible collaborative MIMO in uplink Suitability to High Loading & Throughput per sector
References 1. H. Myung, J. Lim, and D. J. Goodman, Single carrier FDMA (SC-FDMA) for uplink wireless transmission, IEEE VT Magazine, 2006 2. Siavash M. Alamouti, Mobile WiMAX: Vision & Evolution, Intel Corp., January 2008 3. R1-050638, Uplink Multiple Access Scheme, TSG-RAN WG#1, Sophia-Antipolis, 2005 4. J. Zhang, et al., Comparison of the Link Level Performance between OFDMA and SC- FDMA, IEEE ChinaCom, 2006 5. IEEE 80216m-07_037r2, Draft IEEE 802.16m Evaluation Methodology, Dec. 2007 6. IEEE 80216m-07/300, On the Multiple Access Schemes for IEEE 802.16m: Comparison of SC-FDMA and OFDMA, Intel corp, session #52 7. R1-050639, Impact of the transmitter back-off to the uplink range, TSG-RAN WG1 #41bis 8. R1-050743, OFDM PAPR-reduction and associated impact on coverage, TSG-RAN WG1 #42 9. R1-051420, Comparison of system level throughput between SC-FDMA and OFDMA in Evolved UTRA Uplink, TSG-RAN WG1 #43 10. E. Dalhman, S. Parkvall, J. Skold and P. Berning, 3G Evolution: HSPA and LTE for Mobile broadband
Thanks for your attention Q&A