Coverage and Capacity Analysis of mmwave Cellular Systems

Transcription

1 Coverage and Capacity Analysis of mmwave Cellular Systems Robert W. Heath Jr., Ph.D., P.E. Wireless Networking and Communications Group Department of Electrical and Computer Engineering The University of Texas at Austin

2 M MOB (OR) 3 MHz * 32.0 ** 33.0 AERONAUTICA (R) AMATEUR MAR MARITIME 42.0 * MOB MOBI * STANDARD FREQ. AND TIME SI STANDARD FREQ ** 54.0 MO MOBI * * BROADCASTI MARITIME MO MO MOBI Mobile AM AMATEUR AMATEUR 72.0 Mobile 73.0 Mobile M MARITIME MO MOBI 88.0 BR BROADCASTI STANDARD FREQ. AND TIME SIG STANDARD FREQ. MOB AMATEUR Mobile* MOBI MOB BROADCASTI B MARITIME MOBIL MOB RADIO ASTRONOM BROADCASTING BROADCASTING BROADCASTING AMATEUR A AMATEUR STANDARD FREQ. AND TIME SIG STANDARD FREQ. MOBI BROADCASTI MARITIME BROADCASTIN MOB MOBI MARITIME MOBIL STAND. FREQ. & TIME SIG. STANDARD FREQUENCY & TIME SIGN STANDARD FREQ. AMATEUR AM BROADCASTIN MOB (c) 2015 Robert W. Heath Jr. ISM 6.78 ±.015 MHz ISM ±.007 MHz ISM ±.163 MHz AMATEUR Mobile MARITIME MOBIL MARITIME MOBI MOBI AMATEUR STANDARD FREQ. AND TIME SIG STANDARD FREQ. LAND MARITIME MOBI LAND RADIO ASTRONO BROADCASTIN MARITIME MOBI LAND ** AMATEUR AM LAND 30 MHz 300 N Why millimeter wave? LAND LAND LAND LAND LAND Radio Astronomy RADIO ASTRONOMY LAND LAND LAND LAND AMATEUR BROADCASTING (TV CHANNELS 2-4) RADIO ASTRONOMY BROADCASTING (TV CHANNELS 5-6) BROADCASTING (FM RADIO) (R) (R) (R) (R) (R) MET. SAT. (S-E) MOB. SAT. (S-E) RES. (S-E) OPN. (S-E) MET. SAT. (S-E) Mob. Sat. (S-E) RES. (S-E) OPN. (S-E) MET. SAT. (S-E) MOB. SAT. (S-E) RES. (S-E) OPN. (S-E) MET. SAT. (S-E) Mob. Sat. (S-E) RES. (S-E) OPN. (S-E) AMATEUR AMATEUR AMATEUR cellular RADIONAV- LAND LAND LAND MARITIME MARITIME LAND LAND LAND MARITIME MARITIME LAND MARITIME Land Mobile BROADCASTING (TV CHANNELS 7-13) Amateur Fixed Mobile LAND AMATEUR WiFi 30 MHz ISM ±.02 MHz 300 MHz MHz 300 MHz MARITIME 3 GHz LOCATION Amateur LOCATION LOCATION (Ground) STD. FREQ. STD. & TIME FREQ. SIGNAL & TIME SAT. SIGNAL (400.1 SAT. MHz) (400.1 MHz) SAT. (S-E) LOCATION AERO. NAV.(Ground) MET. SAT. (S-E) Space MET. Opn. SAT. (S-E) (S-E) Space Opn. RES. (S-E) (S-E). RES. (S-E) SAT. (S-E). MET. AIDS SAT. (Radiosonde) (S-E) MET. AIDS (Radiosonde) Earth Expl. Earth Expl Sat Satellite SAT. (S-E) Earth Expl. Earth Expl Sat Satellite Met-Satellite EXPL SAT. EXPL MET-SAT. SAT. MET-SAT. (S-E) OPN. MET. AIDS (Radiosonde) (S-E) Earth Expl Sat Met-Satellite EXPL SAT. EXPL MET-SAT. SAT. MET. MET-SAT. AIDS (Radiosonde) MET. AIDS (Radiosonde) METEOROLOGICAL METEOROLOGICAL AIDS (RADIOSONDE) AIDS (RADIOSONDE) RESEARCH RESEARCH (S-S) (S-S) RADIO ASTRONOMY RADIO ASTRONOMY (S-E) RADIOLOCATION RADIOLOCATION Amateur Amateur LAND LAND LAND LAND LAND Meteorological Satellite (S-E) Meteorological Satellite (S-E) LAND LAND LAND LAND LAND LAND LAND LAND LAND LAND (S-E) BROADCASTING BROADCASTING (TV CHANNELS (TV 14 - CHANNELS 20) 14-20) LAND LAND RADIO ASTRONOMY Space Research BROADCASTING (TV CHANNELS 21-36) BROADCASTING (TV CHANNELS 21-36) AERO. RADIONAV. SAT (S-E) RADIOLOCATION LOCATION MARITIME METEOROLOGICAL AIDS MARITIME Amateur LAND LAND RADIO ASTRONOMY RADIO ASTRONOMY Amateur RADIOLOCATION LOCATION Amateur- sat (s-e) TV BROADCASTING TV BROADCASTING 698 BROADCAST BROADCAST 746 BROADCAST BROADCAST RESEARCH Fixed (S-E) ELLITE CY CY TELLITE Met-Satellite OPN. Earth Expl Sat Met-Satellite 4.2 LAND LAND LAND RADIONAV ISM 5.8 ±.075 GHz (S-E) Mobile Satellite (S-E) (S-E) 7.45 BROADCAST BROADCAST MET. (S-E) Mobile Satellite (S-E) Mobile Satellite (S-E) 7.75 (S-E) (S-E) LAND LAND LAND LAND LAND LAND LAND LAND LAND LAND LAND LAND EXPL. (S-E) Mobile Satellite Mobile Satellite (no airborne) Mobile Satellite (no airborne) MET. EXPL. SAT. (S-E) EXPL. (S-E) 8.4 RESEARCH (S-E) (deep space only) LAND LAND 8.45 Amateur Amateur RADIOLOCATION RADIOLOCATION LAND LAND LAND LAND ISM ± 13 MHz ISM ± 13 MHz RESEARCH (S-E) RADIOLOCATION MARITIME 300 MHz 3 GHz UHF (ultra high frequency) spectrum ** ** Amateur SAT (S-E) SAT Fixed Meteorological Aids 9.5 LAND LAND LAND LAND RADIOLOCATION 10.0 Amateur LOCATION Amateur Satellite Amateur RADIOLOCATION RADIO ASTRONOMY EXPL. SAT. RESEARCH EXPL. RESEARCH RADIO ASTRONOMY RADIOLOCATION RADIOLOCATION 1240 (S-E) (S-E) (S-E) 11.7 RADIOLOCATION RADIOLOCATION Amateur Amateur RADIOLOCATION RADIOLOCATION 1390 ** -SAT ** -SAT 1392 ** ** 1395 LAND LAND ASTRONOMY 1400 EXPL SAT EXPL SPA CE SAT RESEARCH RADIO ASTRONOMY RADIO ( Passive) 1427 LAND LAND Fixed (TLM) Fixed (TLM) LAND LAND (TLM) (TLM) (TLM) (TLM) (S-E) Mobile ** 12.2 BROADCASTING 12.7 SPA CE RESEARCH ( Passive) LAND (TLM) -SAT (S-E) -SAT (S-E) (TLM) LAND (TLM) (TLM) RESEARCH (S-E) (Deep Space) ** ** ( ( TELEMETERING) TELEMETERING) SAT. SAT. Mobile ** Mobile ** (Space to Earth) (Space to Earth) Mobile (Aero. TLM) SAT. (Space to Earth) MARITIME SAT. (Space to Earth) RADIONAV. Space Research Standard Freq. and LOCATION Time Signal SAT. Satellite LOCATION Space RADIO Land Mobile Research NAVIGATION SAT. Satellite Space Research Mobile (Aero. TLM) SAT. (Space to Earth) MARITIME SAT. (Space to Earth) 1535 MARITIME MARITIME (space to Earth) (space to Earth) (S-E) (S-E) 1544 (S-E) (S-E) 1545 (R) (space to Earth) Mobile (R) (space to Earth) Satellite (S- E) Mobile Satellite (S- E) Land Mobile Satellite Mobile** 14.4 (Space to Earth) (Space to Earth) (R) (space to Earth) (R) (space to Earth) (R) (space to Earth) (R) (space to Earth) RADIONAV. RADIONAV. (Space to Earth) (Space to Earth) AERO. DET. SAT. RADIO DET. SAT SAT. SAT RADIO DET. SAT. RADIO DET. SAT. RADIO ASTRONOMY SAT. RADIO ASTRONOMY Fixed Mobile SAT. Land Mobile Satellite Fixed Mobile FX SAT. L M Sat Mobile Space Research Fixed Space Research RADIO DET. SAT. RADIO DET. SAT. Mobile SAT. Sat. (S-E) Mobile Sat. (S-E) AERO. RADIONAV. AERO. RADIONAV. AERO. AERO. RADIONAV. AERO. RADIONAV Space Research Mobile RADIO ASTRONOMY RESEARCH EXPL. SAT AERO RADIONAV SAT RADIO ASTRONOMY RADIO ASTRONOMY SAT. SAT. RADIO ASTRONOMY RADIO ASTRONOMY RESEARCH RESEARCH METEOROLOGICAL AIDS (RADIOSONDE) METEOROLOGICAL AIDS (RADIOSONDE) RADIO ASTRONOMY RADIO ASTRONOMY RADIOLOCATION RADIOLOCATION Space Res.(act.) RADIOLOCATION 1670 ** ** METEOROLOGICAL AIDS (Radiosonde) METEOROLOGICAL AIDS (Radiosonde) METEOROLOGICAL (s-e) METEOROLOGICAL (s-e) BCST SAT. FX SAT Fixed Fixed 1710 MET. SAT. (s-e) MET. SAT. (s-e) Radioloc. RADIOLOC. Earth Expl Sat Space Res. (S-E) (S-E) RES. FX SAT (S-E) EXPL. SAT. (S-E) (S-E) 1850 (S-E) SAT. (S-E) FX SAT (S-E) (S-E) STD FREQ. & TIME FX SAT (S-E) SAT (S-E) RES. EXPL. SAT (s-s) RES. EXPL. RES. EXPL. OP. OP SAT. (s-s) SAT. (s-s) (s-s) MOB. FX. MOB. FX S) S) ** ** EXPL. SAT. RAD.AST RES INTER (S-E) (S-E) EXPLORATION SAT. (s-e)(s-s) EXPLORATION OPERATION SAT. (s-e)(s-s) OPERATION RESEARCH (s-e)(s-s) RESEARCH (s-e)(s-s) (LOS) (LOS) (LOS) (LOS) EXPL. SAT. RES. RADIO ASTRONOMY ** RES..(S-E) RES..(S-E) ** Amateur Amateur Amateur RADIOLOCATION Amateur RADIOLOCATION ** ** Mobile Fixed Mobile MOB Fixed FX MOB R- LOC. FX B-SAT 24.0 R- LOC. B-SAT AMATEUR AMATEUR MET. AIDS (Radiosonde) Mobile Radiolocatiolocation Mobile Fixed Radio- BCST- Fixed BCST Mobile Fixed Mobile MOB Fixed FX MOB R- LOC. FX B-SAT R- LOC. B-SAT 2360 RADIOLOCATION RADIOLOCATION Fixed Fixed Amateur AMATEUR AMATEUR AMATEUR ISM ± 50 MHz ISM ± 50 MHz Earth INTER- Expl. Satellite (Active) Amateur RADIOLOCATION INTER- ISM ± GHz Standard Frequency and Time Signal Satellite Earth Exploration Satellite (S-S) RADIODETERMINATION RADIODETERMINATION SAT. (S-E) SAT. (S-E) (S-E) 2500 (S-E) BCST - SAT. BCST ** - SAT. ** FX-SAT (S - E) FX-SAT (S - E) 2655 E-Expl Sat Radio E-Expl Ast Sat Space Radio res. Ast MOB** Space B- res. SAT. MOB** FX FX-SAT B- SAT. FX FX-SAT 2690 RADIO ASTRON. RADIO ASTRON. RESEARCH RESEARCH EXPL SAT EXPL SAT std freq e-e-sat INTER-SAT. & time e-e-sat (s-s) INTER-SAT. Earth Exploration Satellite (S-S) INTER- Earth Exploration Satellite (S-S) 27.5 METEOROLOGICAL AIDS SAT MARITIME MARITIME 3 GHz 3 GHz GHz LOCATION Earth Expl. Satellite (Active) INTER- RADIOLOCATION AMATEUR Amateur METEOROLOGICAL AIDS Standard Frequency and Time Signal Satellite (S-E) Stand. Frequency and Time Signal Satellite (S-E) EXPLORATION SAT. RESEARCH RADIO ASTRONOMY RESEARCH (deep space) INTER- SAT RES. INTER- RADIOLOCATION EXPL. SAT. RE.. RESEARCH (space-to-earth) (S-E) RES. SAT. (S-E) - SAT. EXPL SAT Earth Expl. Sat (s - e) RES. SAT. SAT Mobile Fixed FX-SAT (S-E) BROAD- CASTING BCST SAT. BCST SAT. BROAD- CASTING ** RADIO ASTRONOMY SAT. RADIONAV. MOB. SAT RADIONAV.SAT. AMATEUR AMATEUR FX SAT FX SAT EXPLORATION FI XED RESEARCH RESEARCH EXPLORATION INTER- SAT EXPL-SAT INTER- SAT RES. -ES RES. -ES INTER- SAT EXPLORATION - SAT SAT. RES. INTER RES. EXPLORATION SAT. RESEARCH INTER- SAT LOC. RES.. EXPLORATION SAT. INTER- LOCATION INTER- ** INTER- ** RESEARCH EXPLORATION INTER- RADIO NAVIGATION NAVIGATION AMATEUR AMATEUR RADIOLOC. Amateur RADIOLOC. Amateur Amateur Sat. RADIOLOC. AMATEUR AMATEUR SAT Amateur Satellite Amateur LOCATION (S-E) (S-E) BROAD- CASTING BROAD- CASTING EXPLORATION RESEARCH RADIO ASTRONOMY LOCATION NAVIGATION NAVIGATION RESEARCH EXPL. (S-E) EXPLORATION RESEARCH RADIO ASTRONOMY EXPL SAT. RESEARCH INTER- Amatuer E A R T H EXPL. SAT RES. INTER- SAT. MO- BILE EXPL SAT. RESEARCH INTER- INTER- LOCATION NAVIGATION NAVIGATION AMATEUR AMATEUR Amateur Amateur Satellite LOCATION (S-E) RES. EXPL. SAT. (S-E) (S-E) RES. RADIO ASTRONOMY EXPLORATION INTER- EXPLORATION SAT. INTER- RESEARCH INTER- EXPLORATION RESEARCH RADIO ASTRONOMY INTER- NAVIGATION NAVIGATION EXPLORATION SAT. RES. EXPLORATION RESEARCH RADIO ASTRONOMY (S-E) EXPL. SAT. RES. (S-E) (S-E) Amateur Amateur Satellite LOCATION AMATEUR AMATEUR EXPLORATION RES. NAVIGATION NAVIGATION ASTRONOMY 30 GHz * EXCEPT AERO (R) ** EXCEPT AERO WAVELENGTH BAND DESIGNATIONS ACTIVITIES ISM ±.250 GHz GHz IS DESIGNATED FOR UNLICENSED DEVICES 30 GHz note: log scale so even smaller over here 300 GHz ISM ±.500 GHz 3 x 10 7 m 3 x 10 6 m 3 x 10 5 m 30,000 m 3,000 m 300 m 30 m 3 m 30 cm 3 cm 0.3 cm 0.03 cm 3 x 10 5 Å 3 x 10 4 Å 3 x 10 3 Å 3 x 10 2 Å 3 x 10Å 3Å 3 x 10-1 Å 3 x 10-2 Å 3 x 10-3 Å 3 x 10-4 Å 3 x 10-5 Å 3 x 10-6 Å 3 x 10-7 Å VERY LOW FREQUENCY (VLF) LF MF HF VHF UHF SHF EHF INFRARED VISIBLE ULTRAVIOLET X-RAY GAMMA-RAY COSMIC-RAY Audible Range AM Broadcast FM Broadcast P L S C X Radar Bands Radar Sub-Millimeter Visible Ultraviolet Gamma-ray Cosmic-ray Infra-sonics Sonics Ultra-sonics Microwaves Infrared X-ray FREQUENCY 0 10 Hz 100 Hz 1 khz 10 khz 100 khz 1 MHz 10 MHz 100 MHz 1 GHz 10 GHz 100 GHz 1 THz Hz Hz Hz Hz Hz Hz Hz Hz Hz Hz Hz Hz Hz THE RADIO SPECTRUM 3 khz MAGNIFIED ABOVE 300 GHz Huge amount of spectrum possibly available in mmwave bands ISM ± 1GHz PLEASE NOTE: THE SPACING ALLOTTED THE SERVICES IN THE SPEC- TRUM SEGMENTS SHOWN IS NOT PROPORTIONAL TO THE ACTUAL AMOUNT OF SPECTRUM OCCUPIED. 300 GHz Technology advances make mmwave possible for low cost consumer devices mmwave research is as old as wireless itself, e.g. Bose 1895 and Lebedew

3 The importance of antennas at mmwave millimeter wave band 1.3 GHz 2.1 GHz 7 GHz (unlic) 10 GHz several GHz of spectrum is promising but found in many 28 GHz 37 / 42 GHz 60GHz E-Band to 300 GHz separate bands spatial multiplexing & beamforming just beamforming isotropic radiator mmwave aperture multiple data streams TX RX sub-6ghz aperture Beamforming for antenna gain Spatial multiplexing for spectral efficiency Shu Sun, T. Rappapport, R. W. Heath, Jr., A. Nix, and S. Rangan, `` MIMO for Millimeter Wave Wireless Communications: Beamforming, Spatial Multiplexing, or Both?,'' IEEE Communications Magazine, December

4 Differentiating features of mmwave cellular

5 Antenna scale 64 to 256 elements 4 to 32 elements Large antenna arrays at Tx and Rx Base station Mobile Stations Large antenna arrays result in Large-dimensional precoding/combining matrices High channel estimation, training, and feedback overheads unless smart algorithms exploited Need to design low-complexity precoding and channel estimation algorithms 5

6 Different communication channel bandwidth mmwave noise bandwidth Analog processing Baseband processing UHF noise bandwidth Receiver How to implement equalization? Large channel bandwidth (high noise power, low SNR before beamforming) Implementing random access, channel training and estimation functions is challenging Broadband channels coupled with delay spread Equalization would likely be required at the receiver Hardware constraints may make it difficult to perform equalization entirely in baseband Need new algorithms and architectures for broadband communication 6

7 Hardware constraints 40mW 200mW 200mW 20mW 20mW LNA Analog processing RF Chain ADC Analog processing RF Chain Joint processing ADC Baseband Baseband Processing Precoding Phase shifters Analog processing RF Chain ADC Cost, power, and complexity limit the # of RF chains (high-resolution ADCs) Signal processing can not be done entirely in the baseband Analog beamforming usually uses a network of phase shifters Additional constraints: Constant gain and quantized angles MIMO transceiver DSP algorithms need to incorporate new constraints 7

8 Channel characteristics sub-6 GHz Wifi or Cellular Mi mmwave Wifi mmwave 5G (???) bandwidth 1.4 MHz to 160 MHz 2.16 GHz 100 MHz to 2 GHz # BS or AP 1 to 8 16 to to 256 # antennas at MS 1 or 2 16 to 32 4 to 32 delay spread 100 ns to 10 us 5 to 47 ns 12 to 40 ns angle spread 1 to to 100 up to 50 # clusters 4 to 9 < 4 < 4 orientation sensitivity low medium high small-scale fading Rayleigh Nakagami non-fading or Nakagami large-scale fading distant dependent + distant dependent + distant dependent + shadowing shadowing blockage path loss exponent LOS, 2.5 to 5 NLOS 2 LOS, 3.5 to 4.5 NLOS penetration loss some varies possibly high channel sparsity less more more spatial correlation less more more Some channel characteristics can be leveraged in the signal processing 8

9 Sensitivity to blockages X line-of-sight non-line-of-sight blockage due to people blockage due to buildings Handset User Base station X Blocked by users body hand blocking self-body blocking Need models for these forms of blockage 9

10 MmWave cellular system analysis T. Bai and R. W. Heath Jr., Coverage and rate analysis for millimeter wave cellular networks", IEEE Trans. Wireless Commun, 2015 Also see: T. Bai, A. Alkhateeb and R. W. Heath Jr., Coverage and capacity in millimeter wave cellular networks", IEEE Commun. Mag., vol. 52, no.9, pp , Sep

11 MmWave performance analysis Directional Beamforming (BF) LOS & non-los links Need to incorporate directional beamforming RX and TX communicate via main lobes to achieve array again Steering directions at interfering BSs are random Need to distinguish LOS and NLOS paths Incorporate different characteristics in LOS & NLOS channels Better characterize building blockages Include beamforming + blockage in mmwave cellular analysis 11

12 Accounting for beamforming Exact antenna pattern "Sector" approximation Half-Power BW Back lobe gain Main lobe array gain 7 Antenna Gain m θ M Angel in Rad Sectored antenna pattern approximation Main lobe beamwidth Each base station is marked with a directional antenna Antenna directions of interferers are uniformly distributed Assume perfect beam alignment for desired signal link Use sectored pattern in analysis for simplicity Antenna pattern fully characterized by, M and m 12

13 Incorporating building blockages Snapshots taken from google maps Buildings in some cities, e.g. parts of Boston, form regular grids Use the concept of the LOS probability p(r) Other cities have less regular planning to separate LOS/ NLOS links r A link of length is LOS with probability p(r) LOS probability p(r) is a non-increasing function of the link length r Find the LOS probability based on the certain building models Using stochastic models from random shape theory* Using site-specific maps from geographical information system (GIS) database *T. Bai, R. Vaze, and R. W. Heath Jr., Analysis of blockage effects on urban cellular networks", IEEE Trans. Wireless Commun., vol. 13, no. 9, pp , Sep

14 General mmwave network model LOS BS NLOS BS Typical user Buildings Serving BS NLOS path Typical User LOS path Interfering BSs Use stochastic geometry* to model BS locations as marked PPP Model the steering directions of BSs as independent marks of the point process User connects to the BS with smallest path loss Use distance-dependent LOS probability function Different path loss laws (exponents) for LOS and NLOS paths Assume independent LOS probabilities among links * J. G. Andrews, F. Baccelli, and R. K. Ganti, "A Tractable Approach to Coverage and Rate in Cellular Networks", IEEE TCOM, T. Bai, A. Alkhateeb and R. W. Heath Jr., Coverage and capacity in millimeter wave cellular networks", IEEE Commun. Mag., vol. 52, no.9, pp , Sep p(r) 14

15 Simplified model for dense networks 1 Less than 5% error in coverage LOS BS NLOS BS Typical user Equivalent LOS ball SINR Coverage Probability ISD=200 m, p(t)=e β t, 1/β=200 m LOS ball approximation LOS ball model captures most nearby LOS interferers that dominant the performance in dense networks SINR threshold in db Approximate LOS region by an equivalent LOS ball Theorem 1 can be inefficient to compute due to the general form of p(r) Simplify a general p(r) as a step function by matching its first moment Enable simplified expressions for further performance analysis 15

16 Results on SINR coverage

17 SINR coverage of mmwave cellular Theorem 1 [mmwave SINR Distribution] The SINR coverage probability (CCDF of SINR) in mmwave networks is where the conditional coverage probability by LOS BSs is P c,l (T ) P(SINR >T)=A L P c,l (T )+A N P c,n (T ), X XN L n=1 ( 1) n+1 NL n Z 1 and the conditional coverage probability by NLOS BSs is 0 e n L x L T 2 Q C L MrM n (T,x) V n (T,x) t ˆfL (x)dx, Noise NLOS interf. LOS interf. P c,n (T ) N X N n=1 ( 1) n+1 NN n Z 1 0 e n N x N T 2 W C N MrM n (T,x) Z n (T,x) t ˆfN (x)dx. X Z SINR coverage expressions for general mmwave networks Apply to general building distribution, i.e., LOS probability function p(r) Assume Nakagami fading with different parameters for LOS and NLOS Can be simplified in some special cases of p(r), e.g. step function in the dense network model T. Bai and R. W. Heath Jr., Coverage and rate analysis for millimeter wave cellular networks, IEEE Trans. Wireless Commun,

18 Coverage in dense mmwave networks Theorem 2 [Simplified SINR distribution in dense network] P c (T ) e N X `=1 R ( 1)`+1 Ǹ Z 1 0 exp 2 L b k (` T ā k ) 2 L 2 L ; ` T ā k, ` T ā k s L 2 dt In dense network, SINR largely depends on the relative BS density is defined as the base station density normalized by the LOS region size can be considered as the average number of BSs that are LOS to a user Asymptotic analysis when networks become ultra dense Performance of dense networks limited by LOS interferers, which are typically strong When LOS exponent no larger than 2, asymptotic SINR converges to 0 in probability T. Bai and R. W. Heath Jr., Coverage and rate analysis for millimeter wave cellular networks, IEEE Trans. Wireless Commun, 2015 T. Bai and R. W. Heath Jr., Coverage in dense millimeter wave cellular networks, in Proc. of IEEE Asilomar, Pacific Grove, CA, Nov

19 Coverage gain from large arrays 1 Assume no RX beamforming Signal bandwidth: 500 MHz Avg. ISD: 200 m Avg. LOS range: 1/ =141 m Carrier frequency: 28GHz Tx antenna input power: 30dBm m θ M SINR Coverage Probability Gain from smaller beamwidth (M,m,θ)=(10 db, 10 db, 30 ) (M,m,θ)=(20 db, 10 db, 30 ) (M,m,θ)=(10 db, 10dB, 45 ) Gain from large directivity gain SINR threshold in db SINR coverage benefits from directional beamforming w/ large arrays Larger directivity improve SINR coverage by boosting signal power Small beamwidth reduces the chance of strong interference 19

20 Coverage w/ different BS densities 1 Carrier freq. : 28 GHz Signal Bandwidth: 500 MHz Tx power: 30 dbm Tx directivity gain: 20 db* Tx beamwidth: 30 degree* Rx directivity gain: 10 db Rx beamwidth: 90 degree LOS probability: p(r) =e Avg. LOS range: 1/ =200 m ISD: average inter-site distance r CCDF Noise-limited due to insufficient link budget 0.65 Interference-limited as SINR converges to SIR SINR: ISD=200 m SIR: ISD=200 m SINR: ISD=300 m SIR: ISD=300 m SINR: ISD=400 m SIR: ISD=400 m SINR distribution sensitive to BS density SIR or SINR in db SINR not invariant with BS density due to LOS/NLOS links and noise power From noise-limited to interference-limited when increasing BS density Good coverage achieved when BSs are sufficiently dense * Beamforming 28 GHz from: Z. Pi and F. Khan, "A millimeter-wave massive MIMO system for next generation mobile broadband," In proc. of Asilomar, Nov

21 (c) 2015 Robert W. Heath Jr. SINR coverage comparison 28 GHz strategy: 2 streams enabled by polarization Mode NLOS Optimize # of streams NLOS channel model from [1] No polarization 73 GHz strategy: Beamforming w/ 1 stream CCDF Gain from smaller 73 GHz Due to larger noise 73 GHz 2 GHz (ISD=500 m) 28 GHz (ISD=150 m) 28 GHz (ISD=400 m) 73 GHz (ISD=150 m) 73 GHz (ISD=400 m) SINR in db Comparison of downlink SINR 2 GHz parameters: Signal bandwidth: 50 MHz ISD: 500 m TX power: 46 dbm 4X4 MIMO with ZF receiver 28 GHz parameters: Signal bandwidth: 500 MHz TX power: 30 dbm 8-by-8 UPAs at BSs 4-by-4 UPAs at MSs Each with 4 RF chains 73 GHz parameters: Signal bandwidth: 2 GHz TX power: 30 dbm 20-by-20 at BSs 5-by-5 at Mss Using analog beamforming only (Same aperture size as 28 GHz) Building statistics: LOS range: 70 m (NYU measurement) Perfect CSI at TX and RX [1] M. K. Samimi and T. S. Rappaport, Ultra-wideband statistical channel model for non-line-of-sight millimeter-wave urban channels, Gloabalcomm

22 Results on rate

23 (c) 2015 Robert W. Heath Jr. Rate coverage comparison 28 GHz strategy: 2 streams enabled by polarization Mode NLOS Optimize # of streams NLOS channel model from [1] No polarization 73 GHz strategy: Beamforming w/ 1 stream CCDF GHz (ISD=150 m) 28 GHz (ISD= 400 m) 73 GHz (ISD=150 m) 73 GHz (ISD=400 m) 2 GHz (ISD=500 m) Rate in Mpbs Comparison of per user rate Gain from larger BW Gain from dense BS deployment Gain over conventional cellular system 2 GHz parameters: Signal bandwidth: 50 MHz ISD: 500 m TX power: 46 dbm 4X4 MIMO with ZF receiver 28 GHz parameters: Signal bandwidth: 500 MHz TX power: 30 dbm Using hybrid beamforming: 8-by-8 UPAs at BSs 2-by-2 UPAs at MSs Each with 4 RF chains 73 GHz parameters: Signal bandwidth: 2 GHz TX power: 30 dbm Using analog beamforming: 20-by-20 at BSs 5-by-5 at MSs (Same aperture size as 28 GHz) Building statistics: LOS range: 70 m (NYU measurement) Rate computation: 5 db gap from Shannon Clipping not shown in the plot [1] M. K. Samimi and T. S. Rappaport, Ultra-wideband statistical channel model for non-line-of-sight millimeter-wave urban channels, Gloabalcomm

24 Average rate comparison scenario 5% rate (Mbps) avg rate (Mbps) 2 GHz with 1TX 1RX GHz with 4TX 4RX GHz with sparse BSs (ISD=400 m) 28 GHz with dense BSs (ISD=150 m) 73 GHz with sparse BSs (ISD=400 m) 73 GHz with dense BSs (ISD=150 m) Downlink rate at a typical outdoor user (c) 2015 Robert W. Heath Jr. 2 GHz parameters: Signal bandwidth: 50 MHz ISD: 500 m TX power: 46 dbm MIMO with ZF receiver 28 GHz parameters: Signal bandwidth: 500 MHz TX power: 30 dbm Using hybrid beamforming: 8-by-8 UPAs at BSs 2-by-2 UPAs at MSs Each with 4 RF chains 73 GHz parameters: Signal bandwidth: 2 GHz TX power: 30 dbm Using analog beamforming: 20-by-20 at BSs 5-by-5 at MSs (Same aperture size as 28 GHz) Building statistics: LOS range: 70 m (NYU measurement) Rate computation: 5 db gap from Shannon SINR clipped by 30 db More rate comparison see: Tianyang Bai, Ahmed Alkhateeb, and R. W. Heath, Jr., ``Coverage and Capacity of Millimeter Wave Cellular Networks," IEEE Communications Magazine, Sept

25 Conclusions

26 Going Forward with mmwave Good rates and coverage can be achieved in dense mmwave networks Mmwave is a small cell solution Will magnify gains of densification Many opportunities for further research Analog beamforming algorithms & hybrid beamforming Channel estimation, exploiting sparsity, incorporating robustness Multi-user beamforming algorithms and analysis Microwave-overlaid mmwave systems Going away from cells to a more ad hoc configuration Incorporating mobility questions? 26

27 Select publications (1/4) Performance analysis of mmwave cellular networks [J1,J2, C1,C2] [J1] T. Bai and R. W. Heath Jr., Coverage and rate analysis for millimeter wave cellular networks, IEEE Trans. Wireless Commun., Feb In [J1], we proposed a stochastic geometry network model that incorporated key features of mmwave cellular systems, including directional beamforming and blockage effects. The downlink rate and SINR distributions was then investigated based on the network model. Our analyses showed that mmwave performance is much sensitive to the density of base stations: a denser base station deployment is required to achieve comparable SINR coverage to the conventional cellular networks; the comparable SINR translates to a higher achievable rate, due to the larger bandwidth assumed at mmwave. 27

28 Select publications (2/4) [J2] T. Bai, A. Alkhateeb and R. W. Heath Jr., Coverage and capacity in millimeter wave cellular networks, IEEE Commun. Mag., Sep In [J2], we showed that dense mmwave networks can achieve comparable coverage and significantly higher data rates than the conventional networks. Moreover, sum rate gains can be achieved using more advanced beamforming techniques that allow multiuser transmission. The insights are derived using the framework proposed in [J1]. [C1] T. Bai and R. W. Heath Jr., Coverage in dense millimeter wave cellular networks, in Proc. of IEEE Asilomar, Pacific Grove, CA, Nov We introduced a simplified LOS-ball network model for dense mmwave network analysis. We showed that the performance of dense mmwave networks is largely determined by the average number of LOS base stations that a typical user observes. (Related results also reported in [J1].) 28

29 Select publications (3/4) [C2] T. Bai and R. W. Heath Jr., Analysis of self-body blocking effects in millimeter wave cellular networks, in Proc. of IEEE Asilomar, Pacific Grove, CA, Nov We developed a cone-blocking model to characterize the blocking effect from cellphone users bodies with potential position changes; as human bodies can block mmwave signals causing db attenuation. Based on the network model in [J1], we analyzed the impact of self-body blocking on the SINR coverage and rate under different base station association rules. The results showed that self-body blocking decreases the SINR coverage, and may cause 10% degradation in achievable rates with certain system parameters. 29

30 Select publications (4/4) Modeling building distributions w/ random shape theory [J3] [J3] T. Bai, R. Vaze, and R. W. Heath Jr., Analysis of blockage effects on urban cellular networks, IEEE Trans. Wireless Commun., Sep Leveraging concepts from random shape theory, we modeled the distributions of buildings in urban areas as rectangular Boolean schemes, where the certain of the buildings form a Poisson point process, and their sizes and orientations follow certain distributions. Based on the Boolean scheme model, we showed that the probability that a link is not blocked by any buildings decay exponentially with its length, which matches the LOS probability proposed in 3GPP standard. Furthermore, our analysis on system performance showed that SINR and rate performance in cellular networks can benefit from blockage effects, as buildings may block more interference that often comes from longer links. 30