July 2004 doc IEEE 802 11 040720 r

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 802. 11 n MIMO Link Performance: Some Simulation Results John Ketchum, Bjorn A. Bjerke, and Irina Medvedev Qualcomm, Inc. Submission 1 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Presentation Outline • MIMO system overview • Link simulation overview • Preliminary simulation results: – Throughput and PER performance with 802. 11 n channel B and rate adaptation (CC 67) – PER vs. SNR performance in AWGN (CC 59) • Hardware prototype summary Submission 2 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 MIMO System Overview • • • 2 x 2 – 4 x 4 MIMO-OFDM: enables multiplexing of up to 4 spatial channels OFDM structure identical to 802. 11 a/g (i. e. , 52 subcarriers) 802. 11 a/g constraint length 7, rate ½ convolutional code and punctured rates. 802. 11 a/g QAM modulations + 256 -QAM Additional rates adopted to provide increased spectral efficiency Two approaches to spatial multiplexing: – Eigensteering: • Full CSI operation; Tx & Rx steering based on SVD decomposition per OFDM subcarrier – Spatial spreading: • Partial CSI operation; Rx-only spatial processing • • • Separate coding/interleaving per data stream Interleaving over single OFDM symbol: identical to 802. 11 a/g Closed-loop PHY rate adaptation based on observed Rx SNR Submission 3 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4 MIMO Baseband System Submission 4 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Extended 802. 11 a/g Rate Set Used in Simulations Submission Efficiency (bps/Hz) Data rate (Mbps) Code rate Modulation 0. 5 6 1/2 BPSK 1. 0 12 1/2 QPSK 1. 5 18 3/4 QPSK 2. 0 24 1/2 16 -QAM 2. 5 30 5/8 16 -QAM 3. 0 36 3/4 16 -QAM 3. 5 42 7/12 64 -QAM 4. 0 48 2/3 64 -QAM 4. 5 54 3/4 64 -QAM 5. 0 60 5/6 64 -QAM 6. 0 72 3/4 256 -QAM 7. 0 84 7/8 256 -QAM 5 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Closed-Loop Rate Adaptation • Forward link data rates determined by the STA and communicated to the AP. Rate adaptation algorithm determines the max. data rate supported per spatial stream, based on Rx SNR per subcarrier per stream computed from channel estimates. • Reverse link data rates determined by the AP and communicated to the STA. Rate adaptation algorithm identical to that operating in the STA. Submission 6 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Simulation of Spatial Multiplexing Using Tx & Rx Eigensteering • • • Common MIMO Training Sequence broadcast by AP once every SCAP (Scheduled Access Period) (…, t 0, t 3, …). Forward link (FL) channel coefficients estimated by STA receiver FL Dedicated MIMO Training Sequence (steered) transmitted by AP at t 1=0. 5 ms, immediately followed by FL data PPDU Reverse link (RL) Dedicated MIMO Training Sequence transmitted by STA at t 2=1. 5 ms, immediately followed by RL data PPDU Transmit and receive steering vectors derived from most recent channel estimates Closed-loop rate adaptation: FL and RL data rates determined based on receive SNRs observed in previous frames Submission 7 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Simulation Sequence • • Channel estimation and rate control occurs over three SCAPs (2. 048 ms each) First SCAP: – STA receives Common MIMO Training Sequence • Computes channel estimate • Transmits Dedicated MIMO Training Sequence (steered) – AP receives Dedicated MIMO Training Sequence • Computes estimate of receive & transmit steering vectors • Second SCAP: – AP sends Dedicated MIMO Training Sequence (steered) • STA makes FL rate selection based on received Dedicated MIMO Training Sequence – STA sends Dedicated MIMO Training Sequence (steered) with DL rate selection • AP makes RL rate selection based on received dedicated MIMO Training Sequence • Third SCAP: – AP sends data PPDU based on rate choice in previous SCAP • Includes RL rate selection – STA sends data PPDU SCAP = Scheduled Access Period Submission 8 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Simulation Parameters • 2 x 2 and 4 x 4 system configurations • IEEE 802. 11 n channel models B, D and E • IEEE 802. 11 n impairment models: – Time-domain channel simulator with 5 x oversampling rate (Ts=10 ns) – Rapp nonlinear power amplifier model (IM 1): • Total Tx power = 17 d. Bm; Psat = 25 d. Bm • 2 x 2 backoff = 11 d. B per PA; 4 x 4 backoff = 14 d. B per PA – Carrier frequency offset : -13. 675 PPM (IM 2) – Sampling clock frequency offset: -13. 675 PPM (IM 2) – Phase noise at both transmitter and receiver (IM 4) • 100 channel realizations generated for each SNR point • In each channel realization the Doppler process evolves over three SCAPs to allow simulation of channel estimation, closed-loop rate adaptation and FL/RL data transmission in fading conditions Submission 9 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Simulation Parameters (cont’d) • Stopping criterion: 10 packet errors or 400 packets transmitted per channel realization • Targeted packet error rate performance: mean PER <= 1% • Simulation results: – – – Submission 2 x 2 and 4 x 4 systems Packet size: 1000 B Average throughput vs. SNR Throughput distributions Packet error rate vs. SNR in AWGN (“Fourier channel”) 10 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 2 x 2 Simulation Results: Average throughput vs. SNR Submission 11 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 2 x 2 Throughput Distributions (Forward and Reverse Links) Submission 12 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 2 x 2 Packet Error Rate Distributions (Forward and Reverse Links) Submission 13 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4 Simulation Results: Average throughput vs. SNR Submission 14 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4 Throughput Distributions (Forward and Reverse Links) Submission 15 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4 Packet Error Rate Distributions (Forward and Reverse Links) Submission 16 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Per-Eigenmode Data Rate Distributions • Distribution of data rates picked by the rate selection algorithm in a 4 x 4 system (Forward Link) • Independent data rates per stream (eigenmode) • SNR = 20, 30, 40, 50 d. B • Legend: – 0: Eigenmode does not support transmission – 1: BPSK (0. 5 bps/Hz) – 2, 3: QPSK (1. 0 - 1. 5 bps/Hz) – 4, 5, 6: 16 -QAM (2. 0 - 3. 0 bps/Hz) – 7, 8, 9, 10: 64 -QAM (3. 5 - 5. 0 bps/Hz) – 11, 12: 256 -QAM (6. 0 - 7. 0 bps/Hz) Submission 17 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4, Channel B, SNR = 20 d. B Submission 18 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4, Channel B, SNR = 30 d. B Submission 19 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4, Channel B, SNR = 40 d. B Submission 20 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4, Channel B, SNR = 50 d. B Submission 21 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Packet Error Rate vs. SNR in AWGN • • 2 x 2 and 4 x 4 Fourier matrix channels as specified by CC 59 Perfect synchronization and channel estimation; no impairments Fixed rates (rate adaptation turned off) Simulation results obtained with the following rates: – BPSK, R=1/2 (6 Mbps per stream), – QPSK, R=3/4 (18 Mbps per stream) – 16 -QAM, R=3/4 (36 Mbps per stream) – 64 -QAM, R=3/4 (54 Mbps per stream) – 256 -QAM, R=7/8 (84 Mbps per stream) • Packet size: 1000 B Submission 22 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 2 x 2 PER vs. SNR Submission 23 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 4 x 4 PER vs. SNR Submission 24 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Acquisition Performance • • Probability of detection Frequency estimate error variance Frequency estimate error distributions 40 ppm timing and frequency offset @5. 25 GHz • 1000 channel realizations, 20 trials per channel realization Submission 25 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Short Training Sequence Detection Performance in TGn Channels Submission 26 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Frequency Estimation Performance in TGn Channels Submission 27 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Frequency Error CDF SNR = 10 d. B Submission 28 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 MIMO WLAN Prototype Submission 29 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Qualcomm MIMO WLAN Prototype • • Fully operational hardware prototype of 802. 11 n MIMO WLAN modem Supports four antennas for Tx/Rx Processing implemented in two Virtex-II Xilinx FPGAs FPGA board has four channels that are digitized using four 12 -bit A/D converters • All signal processing is real-time – IF sampled at 80 MHz is converted to complex baseband at 20 MHz sampling rate – FFT and spatial processing runs at peak rate of 250 k. Hz – Multiple parallel convolutional decoders run at 80 MHz clock to provide highspeed decoding – Fully operational hardware channel estimation and tracking • Supports stationary and low-mobility operation • Closed-loop rate control performs real-time rate adjustments • Performance matches results of detailed simulations • Operates in PCS band (1930 – 1950 MHz) – 0 d. Bm per antenna Tx power to limit interference to PCS users per experimental license Submission 30 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 MIMO WLAN Prototype Submission 31 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 MIMO WLAN Prototype Submission 32 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 MIMO WLAN Measured Performance • Measurements made in Qualcomm’s Concord MA offices – Renovated mill building with brick outer walls, sheet-rock on metal studs inner walls – 41. 2 m x 16. 8 m, 3. 2 m high floor-to-ceiling walls • Measured values indicate error-free physical-layer data rates achieved – All rates reflect communications between AP and stations • Measurements made at 1940 MHz • Measurement results extrapolated to 2. 4 GHz and 5. 25 GHz unlicensed bands – Data rates scaled to reflect operation in SNRs were modified to reflect 17 d. Bm total Tx power, and increased path loss at 2. 4 GHz (1. 6 d. B) and 5. 25 GHz (8. 4 d. B) – Resulting median/minimum physical layer data rates: • 2. 4 GHz: 310/198 Mbps • 5. 25 GHz: 225/120 Mbps Submission 33 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Measured PHY Data Rates Submission 34 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Extrapolated PHY Data Rates Submission 35 John Ketchum, et al. , Qualcomm Incorporated

July 2004 doc. : IEEE 802. 11 -04/0720 r 0 Summary • Eigenvector steered mode supports high throughput operation in 2 x 2 and 4 x 4 configurations • Stable wideband spatial channels synthesized from eigenmodes easily support 256 QAM under full PHY impairments • High throughput eigenvector steering operation proven in hardware prototype Submission 36 John Ketchum, et al. , Qualcomm Incorporated