19, Yangjae-daero 11gil, Seocho-gu, Seoul 137-130, Korea Month Year doc.: IEEE 802.11-yy/xxxxr0 September 2015 1024 QAM Proposal Date: 2015-09-14 Authors: Name Affiliation Address Phone Email Eunsung Park LG Electronics 19, Yangjae-daero 11gil, Seocho-gu, Seoul 137-130, Korea   esung.park@lge.com Jinsoo Choi js.choi@lge.com Jinyoung Chun jiny.chun@lge.com Dongguk Lim dongguk.lim@lge.com Jinmin Kim jinmin1230.kim@lge.com Kiseon Ryu kiseon.ryu@lge.com Jeongki Kim jeongki.kim@lge.com Suhwook Kim suhwook.kim@lge.com Hyeyoung Choi hy0117.choi@lge.com HanGyu Cho hg.cho@lge.com Eunsung Park, LG Electronics John Doe, Some Company

2111 NE 25th Ave, Hillsboro OR 97124, USA September 2015 Authors (continued) Name Affiliation Address Phone Email Ron Porat Broadcom   rporat@broadcom.com Sriram Venkateswaran mfischer@broadcom.com Matthew Fischer Leo Montreuil Andrew Blanksby Vinko Erceg Robert Stacey Intel 2111 NE 25th Ave, Hillsboro OR 97124, USA     +1-503-724-893   robert.stacey@intel.com Eldad Perahia eldad.perahia@intel.com Shahrnaz Azizi shahrnaz.azizi@intel.com Po-Kai Huang po-kai.huang@intel.com Qinghua Li quinghua.li@intel.com Xiaogang Chen xiaogang.c.chen@intel.com Chitto Ghosh chittabrata.ghosh@intel.com Laurent cariou laurent.cariou@intel.com Rongzhen Yang rongzhen.yang@intel.com Assaf Gurevitz assaf.gurevitz@intel.com Eunsung Park, LG Electronics

5488 Marvell Lane, Santa Clara, CA, 95054 September 2015 Authors (continued) Name Affiliation Address Phone Email Hongyuan Zhang Marvell 5488 Marvell Lane, Santa Clara, CA, 95054 408-222-2500 hongyuan@marvell.com Yakun Sun yakunsun@marvell.com Lei Wang Leileiw@marvell.com Liwen Chu liwenchu@marvell.com Jinjing Jiang jinjing@marvell.com Yan Zhang yzhang@marvell.com Rui Cao ruicao@marvell.com Jie Huang jiehuang@marvell.com Sudhir Srinivasa sudhirs@marvell.com Saga Tamhane sagar@marvell.com Mao Yu my@marvel..com Edward Au edwardau@marvell.com Hui-Ling Lou hlou@marvell.com Eunsung Park, LG Electronics

Authors (continued) September 2015 Albert Van Zelst Qualcomm Name Affiliation Address Phone Email Albert Van Zelst Qualcomm Straatweg 66-S Breukelen, 3621 BR Netherlands   allert@qti.qualcomm.com Alfred Asterjadhi 5775 Morehouse Dr. San Diego, CA, USA aasterja@qti.qualcomm.com Bin Tian btian@qti.qualcomm.com Carlos Aldana 1700 Technology Drive San Jose, CA 95110, USA caldana@qca.qualcomm.com George Cherian gcherian@qti.qualcomm.com Gwendolyn Barriac gbarriac@qti.qualcomm.com Hemanth Sampath hsampath@qti.qualcomm.com Menzo Wentink mwentink@qti.qualcomm.com Richard Van Nee rvannee@qti.qualcomm.com Rolf De Vegt rolfv@qca.qualcomm.com Sameer Vermani svverman@qti.qualcomm.com Simone Merlin smerlin@qti.qualcomm.com Tevfik Yucek   tyucek@qca.qualcomm.com VK Jones vkjones@qca.qualcomm.com Youhan Kim youhank@qca.qualcomm.com Eunsung Park, LG Electronics

Authors (continued) September 2015 James Yee Mediatek Name Affiliation Address Phone Email James Yee Mediatek No. 1 Dusing 1st Road, Hsinchu, Taiwan +886-3-567-0766  james.yee@mediatek.com Alan Jauh   alan.jauh@mediatek.com Chingwa Hu chinghwa.yu@mediatek.com Frank Hsu frank.hsu@mediatek.com Thomas Pare USA 2860 Junction Ave, San Jose, CA 95134, USA +1-408-526-1899 thomas.pare@mediatek.com ChaoChun Wang chaochun.wang@mediatek.com James Wang james.wang@mediatek.com Jianhan Liu Jianhan.Liu@mediatek.com Tianyu Wu tianyu.wu@mediatek.com Russell Huang russell.huang@mediatek.com Joonsuk Kim Apple    joonsuk@apple.com Aon Mujtaba   mujtaba@apple.com Guoqing Li guoqing_li@apple.com Eric Wong ericwong@apple.com  Chris Hartman chartman@apple.com Eunsung Park, LG Electronics

Authors (continued) September 2015 Phillip Barber Huawei Peter Loc Name Affiliation Address Phone Email Phillip Barber Huawei The Lone Star State, TX   pbarber@broadbandmobiletech.com Peter Loc peterloc@iwirelesstech.com Le Liu F1-17, Huawei Base, Bantian, Shenzhen +86-18601656691 liule@huawei.com Jun Luo 5B-N8, No.2222 Xinjinqiao Road, Pudong, Shanghai jun.l@huawei.com Yi Luo +86-18665891036 Roy.luoyi@huawei.com Yingpei Lin linyingpei@huawei.com Jiyong Pang pangjiyong@huawei.com Zhigang Rong 10180 Telesis Court, Suite 365, San Diego, CA  92121 NA zhigang.rong@huawei.com Rob Sun 303 Terry Fox, Suite 400 Kanata, Ottawa, Canada Rob.Sun@huawei.com David X. Yang david.yangxun@huawei.com Yunsong Yang yangyunsong@huawei.com Zhou Lan F1-17, Huawei Base, Bantian, SHenzhen +86-18565826350 Lanzhou1@huawei.com Junghoon Suh Junghoon.Suh@huawei.com Jiayin Zhang zhangjiayin@huawei.com Eunsung Park, LG Electronics

Authors (continued) September 2015 Fei Tong Samsung Hyunjeong Kang Name Affiliation Address Phone Email Fei Tong Samsung Innovation Park, Cambridge CB4 0DS (U.K.) +44 1223 434633 f.tong@samsung.com Hyunjeong Kang Maetan 3-dong; Yongtong-Gu Suwon; South Korea +82-31-279-9028 hyunjeong.kang@samsung.com Kaushik Josiam 1301, E. Lookout Dr, Richardson TX 75070 (972) 761 7437 k.josiam@samsung.com Mark Rison +44 1223 434600 m.rison@samsung.com Rakesh Taori (972) 761 7470 rakesh.taori@samsung.com Sanghyun Chang +82-10-8864-1751 s29.chang@samsung.com Yasushi Takatori NTT 1-1 Hikari-no-oka, Yokosuka, Kanagawa 239-0847 Japan   takatori.yasushi@lab.ntt.co.jp Yasuhiko Inoue inoue.yasuhiko@lab.ntt.co.jp Yusuke Asai asai.yusuke@lab.ntt.co.jp Koichi Ishihara ishihara.koichi@lab.ntt.co.jp Junichi Iwatani Iwatani.junichi@lab.ntt.co.jp Shoko Shinohara Shinohara.shoko@lab.ntt.co.jp Akira Yamada NTT DOCOMO 3-6, Hikarinooka, Yokosuka-shi, Kanagawa, 239-8536, Japan yamadaakira@nttdocomo.com Fujio Watanabe 3240 Hillview Ave, Palo Alto, CA 94304 watanabe@docomoinnovations.com Haralabos Papadopoulos hpapadopoulos@docomoinnovations.com Eunsung Park, LG Electronics

#9 Wuxingduan, Xifeng Rd., Xi'an, China September 2015 Authors (continued) Name Affiliation Address Phone Email Yuichi Morioka Sony Corporation 1-7-1 Konan  Minato-ku, Tokyo 108-0075, Japan    Yuichi.Morioka@jp.sony.com Masahito Mori Masahito.Mori@jp.sony.com Yusuke Tanaka YusukeC.Tanaka@jp.sony.com Kazuyuki Sakoda Kazuyuki.Sakoda@am.sony.com William Carney William.Carney@am.sony.com Thomas Derham Orange   thomas.derham@orange.com Bo Sun ZTE #9 Wuxingduan, Xifeng Rd., Xi'an, China   sun.bo1@zte.com.cn Kaiying Lv lv.kaiying@zte.com.cn Yonggang Fang yfang@ztetx.com Ke Yao yao.ke5@zte.com.cn Weimin Xing xing.weimin@zte.com.cn Brian Hart Cisco Systems 170 W Tasman Dr, San Jose, CA 95134 brianh@cisco.com Pooya Monajemi pmonajem@cisco.com Eunsung Park, LG Electronics

Introduction This contribution proposes to use 1024 QAM in 11ax September 2015 Introduction This contribution proposes to use 1024 QAM in 11ax First part talks about the importance of peak throughput enhancement of 11ax for marketing point of view Second part provides system level simulations to show average throughput enhancement by 1024QAM Finally, we discuss feasibility of 1024QAM and propose to use it optionally in 11ax Eunsung Park, LG Electronics

How to appeal 11ax in the market? September 2015 How to appeal 11ax in the market? We hope 11ax becomes a big success in market But, ‘average throughput enhancement of 4X’ may not be so appealing to market even if it is very important technically. 1024 QAM provides a good marketing point on 11ax based on the followings: WLAN finally achieves 10Gbps (Multi-gigabit)! New frame structure provides around 15% enhancement of peak throughput (4X FFT with 0.4us CP, More guard/pilot tones used as data) 1024 QAM provides additional 25% gain, which results in overall 44% gain (6.933Gbps * 1.44 = 9.98Gbps!) The first wireless technology to use 1K QAM1) Keep superior to other standards in terms of modulation technology (LTE decided to use 256QAM) Eunsung Park, LG Electronics

How to meet the technical target of 11ax? September 2015 How to meet the technical target of 11ax? Based on our system-level simulation, we have the following observations. Simulation parameters are described in Appendix A Observation 1. In most indoor scenarios, 1024 QAM MCS levels are selected with very high probability (see page 5) We additionally used MCS 10 (1024QAM with 3/4 code rate) and MCS 11 ( 1024QAM with 5/6 code rate) Observation 2. 1024 QAM provides average throughput gain over 20% in most indoor scenarios (see page 6) Eunsung Park, LG Electronics

MCS Selection Probability with 1024QAM September 2015 MCS Selection Probability with 1024QAM We simulated SISO, space time block code (STBC), and spatial multiplexing (SM) in each simulation scenario [2] Residential Enterprise Indoor hotspot Outdoor SISO STBC SM Eunsung Park, LG Electronics

Average Throughput Gain with 1024QAM September 2015 Average Throughput Gain with 1024QAM Average Throughput Gain by including 1024 QAM [Unit : %] Residential Enterprise Indoor hotspot Outdoor DL UL SISO 22 18 25 21 20 11 1 STBC 24 12 2 SM 13 14 3 5 Eunsung Park, LG Electronics

Discussion on Feasibility (1/2) September 2015 Discussion on Feasibility (1/2) For 1024 QAM application, we need to discuss several issues such as Required minimum EVM Non-linearity of power amplifier (PA) Quantization error of analog-to-digital converter (ADC) Phase noise and I/Q imbalance of local oscillator (LO) Residual CFO effect Constellation design, interleaver design, etc Eunsung Park, LG Electronics

Discussion on Feasibility (2/2) September 2015 Discussion on Feasibility (2/2) EVM of -35dBc or even -34dBc give reasonable performance (See Appendix B) Note that even in 802.11ac MU-MIMO the AP already has to improve its Tx EVM level to about -38dBc, so 1024 QAM should not impose a real new requirement Assuming future modems will include sophisticated algorithms that will combat Tx EVM sources (phase noise, PA nonlinearities, IQ imbalance), the Tx EVM requirement can be further relieved Current ADC technology seems to make 1024 QAM feasible ADC needs 12 bits for quantization at least when using 1024 QAM and it is already possible to implement CFO may not cause any problem with a proper pilot design Although a high modulation level such as 1024 QAM is vulnerable to CFO, properly designed pilot can reliably estimate and compensate residual CFO Eunsung Park, LG Electronics

September 2015 Conclusion We have showed that using 1024 QAM in 11ax can be appeal to the market and also help to guarantee the technical target of 11ax by providing peak data rate increase and average throughput enhancement Also, we have checked the feasibility of 1024 QAM and the current technology seems feasible Thus, we propose to use 1024 QAM in 11ax Eunsung Park, LG Electronics

September 2015 Straw poll 1 Do you agree to use 1024 QAM as an optional feature for SU and MU using the resource units equal to or larger than 242 tones in 11ax? Y/N/A Eunsung Park, LG Electronics

September 2015 References [1] IEEE 802.11-14/0624r0 - Investigation on 1024QAM Feasibility in 11ax [2] IEEE 802.11-14/0980r12 - Simulation Scenarios Eunsung Park, LG Electronics

September 2015 Appendices Eunsung Park, LG Electronics

A. Simulation Parameters September 2015 A. Simulation Parameters Simulation scenario Scenario1 Scenario2 Scenario3 Scenario4 AP/STA power [dBm] 23 / 17 24 / 21 17 / 15 30 / 15 # of users per BSS 5 32 30 50 Run time Initial: 10s, simulation: 10s, drop: 5 BW 40MHz (128 FFT) Data size 1ms TXOP less fixed overhead (RTS/CTS off) Normal overhead: SIFS + ACK + 2*PLCP header GI Long (0.8 us) # of AP/STA antennas Rank1: SISO(1,1)/Alamouti(2,2), Rank2: SM(2,2) (interference model: AWGN) Max # of retries Scenario1: 4, other scenarios: 10 DL & UL traffic Full buffer (DL & UL ratio: based on PHY system simulation in [2]) CCA level Preamble detection: -82dBm (both AP and STA) Energy detection: -62dBm (both AP and STA) Channel Time-varying Target PER 0.1 PHY abstraction Capacity based method MCS selection Genie method Eunsung Park, LG Electronics

B. Effect of TX EVM on RX performance September 2015 B. Effect of TX EVM on RX performance 4 Tx 2 Rx MIMO 2 streams, 1024 QAM rate 5/6, no channel estimation loss, beamforming SVD, ZF equalizer Tx EVM level (dBc) D channel low corr (SNR) D channel high corr (SNR) No Tx EVM 31.2 35. 9 -38 31.6 36.1 -35 32.1 (0.9 dB loss) 36.6 (0.7 dB loss) -34 32.4 37 -32 34 40 Eunsung Park, LG Electronics

C. Link Curve MCS 11 requires about 32dB SNR for PER of 10%. September 2015 C. Link Curve MCS 11 requires about 32dB SNR for PER of 10%. No CFO, no TO channel estimation w/ LS Eunsung Park, LG Electronics

September 2015 Back up Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.7, 4 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.7, 4 Tx antennas, 2 Rx antennas Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.7, 4 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.7, 4 Tx antennas, 2 Rx antennas EVM = 35 (1024 QAM), 32 (256 QAM) Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.3, 4 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.3, 4 Tx antennas, 2 Rx antennas Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.3, 4 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.3, 4 Tx antennas, 2 Rx antennas EVM = 35 (1024 QAM), 32 (256 QAM) Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.7, 3 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.7, 3 Tx antennas, 2 Rx antennas Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.7, 3 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.7, 3 Tx antennas, 2 Rx antennas EVM = 35 (1024 QAM), 32 (256 QAM) Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.3, 3 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.3, 3 Tx antennas, 2 Rx antennas Eunsung Park, LG Electronics

Simulation Results Channel D, corr=0.3, 3 Tx antennas, 2 Rx antennas September 2015 Simulation Results Channel D, corr=0.3, 3 Tx antennas, 2 Rx antennas EVM = 35 (1024 QAM), 32 (256 QAM) Eunsung Park, LG Electronics