2003-03-06 IEEE C802.20-03/12
1
Project IEEE 802.20 Working Group on Mobile Broadband Wireless Access <http://grouper.ieee.org/groups/802/20 >
Title Antenna Arrays for MBWA: Overview and Field Experiments
Date Submitted
2003-03-06
Source(s) Frederick W. Vook Motorola Labs Communication Systems Research Laboratory 1301 E. Algonquin Road, IL02-2912 Schaumburg, IL 60196
Email: [email protected]
Re: MBWA ECSG Call for Contributions
Abstract This submission presents an overview of antenna array technologies for mobile broadband wireless systems and presents recent field results from a 2x2 OFDM experimental testbed.
Purpose For informational use only
Notice This document has been prepared to assist the IEEE 802.20 Working Group. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) 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 802.20.
Patent Policy
The contributor is familiar with IEEE patent policy, as outlined in Section 6.3 of the IEEE-SA Standards Board Operations Manual <http://standards.ieee.org/guides/opman/sect6.html#6.3> and in Understanding Patent Issues During IEEE Standards Development <http://standards.ieee.org/board/pat/guide.html>.
Antenna Arrays for MBWA: Antenna Arrays for MBWA: Overview and Field Experiments Overview and Field Experiments
IEEE 802.20March 10-14, 2003
2 of 16March 2003 IEEE 802.20 MBWA
Motivation – Why Antenna Arrays? Motivation Motivation –– Why Antenna Arrays? Why Antenna Arrays?
• Ultimate Goal: Increase Capacity and Reliability
• Capabilities:– Coherent beamforming gain– Space-Diversity Exploitation– Interference Suppression
• Increase capacity via:– Enable higher order modulation– Enable smaller re-use factors
• Multiply capacity: spatial multiplexing– Spatial Division Multiple Access (SDMA)– Multiple Input Multiple Output (MIMO)
• Benefits:– Improve Link Quality– Improve Coverage Reliability– Enhance Range
• Current standards are incorporating Antenna Array techniques– 3G CDMA: Space Time Transmit Diversity (STTD/Alamouti), Transmit
Adaptive Arrays (TXAA), with MIMO on the horizon• Many different techniques for exploiting multiple antennas• For best results, AAs must be matched to the entire system• Gains are environmentally dependent
3 of 16March 2003 IEEE 802.20 MBWA
Challenges in Broadband Mobile SystemsChallenges in Broadband Mobile SystemsChallenges in Broadband Mobile Systems
• Mobile Broadband is a challenging environment for Antenna Arrays– High mobility: causes rapid variations across the time-dimension– Multipath delay spread: causes severe frequency-selective fading– Multipath angular spread: causes significant variations in the spatial channel
responses of the incident signals– For best performance, the Rx & Tx algorithms must accurately track all
dimensions of the channel responses (time, frequency, and space)
0 5 10-50
-40
-30
-20
-10
0
de lay (microsec)
mag
nitu
de (d
B)
pExample 2:Example 1:
4 of 16March 2003 IEEE 802.20 MBWA
Frequency-Domain Transmit Array ProcessingFrequencyFrequency--Domain Transmit Array ProcessingDomain Transmit Array Processing
• Orthogonal Frequency Division Multiplexing (OFDM)• Cyclic-Prefix Single Carrier• Better complexity & performance vs. time-domain approaches in
broadband high delay spread channels
N-point IFFT
N-point IFFT
. . .
N-point FFT
. . .
Array ProcessingThis FFT is not needed in OFDM
Add prefix
Add prefix
General Multiple Antenna Transmitter
Cyclic Prefix
Time
Data Portion(Length N)
. . .. . .
Output of IFFT (OFDM or SOFDM)Or block of symbols (CP-single carrier)
Normally a copy of the last Np samples of the data portionTransmission Format:
5 of 16March 2003 IEEE 802.20 MBWA
Frequency-Domain Receive Array ProcessingFrequencyFrequency--Domain Receive Array ProcessingDomain Receive Array Processing
• Max Ratio Diversity Combining– Optimize for Maximum S/N– Beamforming gain over noise– Diversity gain in faded channels
• Optimal Combining– Optimize for Maximum S/(I+N)– Tradeoff between Beamforming &
Diversity Gain and Interference Suppression
N-point FFT
N-point FFT
. . . N-point IFFT
. . .
Equalization W
eighting and C
ombining
This IFFT is not needed for OFDM
Remove prefix
Remove prefix
General Multiple Antenna Receiver
10
20
30
40
30
210
60
240
90
270
120
300
150
330
180 0
Active Interference Suppression with an 8-Element circular array
Angular Array Response on one OFDM subcarrier (simulated)
6 of 16March 2003 IEEE 802.20 MBWA
x1(k)
x2(k)
v1(k)
v2(k)
x (k)Mtx
vM (k)tx
s(k)InputSymbolStream
o o o
y1(k)
y2(k)
y (k)Mrx
w1(k)
w2(k)
wM (k)rx
Σz(k)
OutputSymbolStream
o o o
Transmit and Receive Array Processing
Common Transmit Array TechniquesCommon Transmit Array TechniquesCommon Transmit Array Techniques
• Alamouti Transmit Diversity– Multiplex two QAM / PSK symbols
onto two antennas over two symbol intervals.
– TX diversity gain with no channel knowledge at transmitter
– Incorporated in 3G-CDMA (STTD)– Easily applied to OFDM
• Apply across bauds or across subcarriers
– Easily applied to CP-Single Carrier• Time-Reversal & Conjugation trick
OFDMTx.
symbols
OFDMTx.
OFDMTx.
Encoder
Interleaver
Group &Modulate
TX ArrayProcessor
BICM encode
Space-Time Coding
symbolsOFDMRx.
OFDMRx.
RX Array Processor
Bit de-interleaver
Decoder
BICM decode
Computebit metric
• Transmit Adaptive Beamforming– Direction-based tracking of subscriber– Focus Tx energy towards subscriber– Reduces interference to other cells– Limited Tx diversity gain in fading
• Transmit Adaptive Array (TXAA)– Incorporated into 3G-CDMA– A diversity-spaced array provides both
beamforming & diversity gains• FDD: feedback• TDD: reuse uplink information
– Gains diminish with inaccurate channel information in mobile channels
7 of 16March 2003 IEEE 802.20 MBWA
Spatial Division Multiple Access (SDMA)Spatial Division Multiple Access (SDMA)Spatial Division Multiple Access (SDMA)
• Base station communicates with multiple subscriber devices on the same time-frequency resources simultaneously
• Multiply capacity by serving multiple users simultaneously• Receive SDMA relies on multi-user channel estimation and tracking
along with baseband array combining algorithms• Transmit SDMA may be difficult to implement in fast-moving
broadband channels– Need precise channel information at the transmit array to eliminate cross-
talk between the spatial channels
Multiple user signals to be transmitted on same time-frequency resourcesMultiple user signals received on same
time-frequency resources
TX SDMARX SDMA
y1(k)
y2(k)
y (k)Mrx
z1(k)
w11(k)
w21(k)
wM 1(k)rx
w12(k)
w22(k)
wM 2(k)rx
Σ Σ
z2(k)
Σo o o
wM N (k)rx s
w2N (k)s
w1N (k)s
zN (k)s
x1(k)
x (k)Ns
x2(k)
x1(k)
x2(k)
x (k)Mtx
v11(k)
v21(k)
vM 1(k)tx
s1(k)
Σ
Σ
Σ
s2(k) sNs(k)
vM 2(k)tx
v22(k)
v12(k)
vM N (k)tx
v2N (k)
v1N (k)
s
s
s
y1(k)
y2(k)
y (k)Ns
Multiple user signals transmitted on same time-frequency resources
8 of 16March 2003 IEEE 802.20 MBWA
Multiple Input Multiple Output (MIMO)Multiple Input Multiple Output (MIMO)Multiple Input Multiple Output (MIMO)
• Transmitting one or more data streams over multiple spatial channels between a single TX and RX device
• Advantage:– Vastly increased theoretical capacity vs single-stream/antenna methods– Practical view: Form multiple spatial channels each using a small
modulation & coding rate rather than using a single spatial channel having a large modulation & coding rate
• Disadvantage:– MIMO methods need sufficient angular multipath scattering so that the
transmit antennas are “spatially separable”– MIMO methods fail when channel matrix has high levels of correlation
Open-Loop MIMO Closed-Loop MIMO
x1(k)
x2(k)
x (k)Mtx
v11(k)
v21(k)
vM 1(k)tx
s1(k)
Σ
Σ
Σ
s2(k) sNs(k)
vM 2(k)tx
v22(k)
v12(k)
vM N (k)tx
v2N (k)
v1N (k)
s
s
s
y1(k)
y2(k)
y (k)Mrx
z1(k)
w11(k)
w21(k)
wM 1(k)rx
w12(k)
w22(k)
wM 2(k)rx
Σ Σ
z2(k)
Σo o o
wM N (k)rx s
w2N (k)s
w1N (k)s
zN (k)s
Multiple Input Data Streams Multiple Output Data Streams
x1(k)
x2(k)
x (k)Mtx
s1(k)
s2(k)
sNs(k)
.
.
.
Multiple Input Data Streams from a single user
y1(k)
y2(k)
y (k)Mrx
z1(k)
w11(k)
w21(k)
wM 1(k)rx
w12(k)
w22(k)
wM 2(k)rx
Σ Σ
z2(k)
Σo o o
wM N (k)rx s
w2N (k)s
w1N (k)s
zN (k)s
Multiple Output Data Streams
9 of 16March 2003 IEEE 802.20 MBWA
Field Data Collection DescriptionField Data Collection DescriptionField Data Collection Description
Test TruckBase Site Antennas
3.675 GHz carrier20 MHz channel BW
Two identical & independent Rx5 dBi omni antennas, spaced ~9.3 λ (~75cm)Synchronized to GPS and received signalTime & Frequency domain data720 snapshots of 9 MBytes per hour, 6.4GB/h
6 sectors, 2 antennas/sectorLocated on top of 6-story building5 λ antenna spacing (~41 cm)18 dBi antenna gain (80º beamwidth)
10 of 16March 2003 IEEE 802.20 MBWA
Theoretical 2x2 Capacity Gain over 1x1 Based on Measured ChannelsTheoretical 2x2 Capacity Gain over 1x1 Theoretical 2x2 Capacity Gain over 1x1 Based on Measured ChannelsBased on Measured Channels
• Dependency on DOA spread– Directions of Arrival measured with
synthetic aperture method (To appear: Krauss, et al., VTC-2003-Spring, April 2003)
– Higher DOA spreads correspond to higher capacity gain over 1x1
• Dependency on Delay Spread– Higher Delay Spreads tend to
correspond to rich scattering conditions
1 1.2 1.4 1.6 1.8 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Capacity gain
Pro
babi
lity
0 to 0.5 (1183)0.5 to 1.5 (1211)1.5 to 10 (442)All (2836)
1 1.2 1.4 1.6 1.8 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Capacity gain
Pro
babi
lity
0 to 22.5 (827)22.5 to 45 (1215)45 to 90 (794)All (2836)
• Multi-antenna Shannon capacity formula (see Foschini, Teletar, etc.)• Using measured frequency-domain channel matrices• CDF of open-loop 2x2 capacity gain over 1x1
Increasing DOA SpreadALL
Increasing Delay SpreadALL
DOA Spread Bins Delay Spread Bins
11 of 16March 2003 IEEE 802.20 MBWA
Alamouti vs MIMO Alamouti Alamouti vs vs MIMO MIMO
• QAM Constellations & Uncoded BER (Off-line)
• Non-linear receiver processing and coding can help achieve the higher capacity benefits of MIMO transmission
• For MIMO to perform well, need good spatial conditioning
• “Alamouti” scheme hasboth transmit and receive diversity (2x2)
• 2x2 MIMO processing:- linear MMSE
equalization- offers 2x the data rate
12 of 16March 2003 IEEE 802.20 MBWA
Spatial Conditioning: Open-Loop 2x2 MIMO PerformanceSpatial Conditioning: Spatial Conditioning: OpenOpen--Loop 2x2 MIMO PerformanceLoop 2x2 MIMO Performance
• The average reciprocal condition number (0< κ-1 <1) gives a rough sense of how well 2x2 MIMO will work
κ-1 =1: The spatial Tx signatures are orthogonal and equal magnitudeκ-1 =0: Singular 2x2 channel matrix, can’t separate the two Tx streams
• Although κ-1 never reaches unity, the 64-QAM performance indicates that there is enough spatial separation much of the time to support MIMO operation in this environment
• Field observation: κ-1 >.3 indicates good spatial conditioning, κ-1 <.2 indicates poor spatial conditioning
• In this environment: ~30% of the time the channel exhibits good spatial conditioning (i.e. the channel is suitable for MIMO operation)
κ-1
13 of 16March 2003 IEEE 802.20 MBWA
Pushing the Limits of 2x2 MIMO (300 Mbps)Pushing the Limits of 2x2 MIMO (300 Mbps)Pushing the Limits of 2x2 MIMO (300 Mbps)
• BER vs. position for Uncoded 2-stream High Order-QAM w/ MIMO– Optimal channel estimation, 10dB excess Tx power, no interference– Receive antennas mounted on top of the test truck– 300 Mb/second channel data rate (18.8 MHz Bandwidth)
• Low BER locations indicate sufficient multipath scattering.
• Although this is an idealized case, it does show that MIMO detection is worth further consideration
14 of 16March 2003 IEEE 802.20 MBWA
Alamouti vs. MIMO in measured 2x2 channelsAlamouti vs. MIMO in measured 2x2 channelsAlamouti vs. MIMO in measured 2x2 channels
• Decoded FER performance for different ranges of the reciprocal condition number of the matrix channel response
• Alamouti (2-2) performance fairly constant with the condition number• MIMO (2-2) performance degrades significantly in poorly conditioned
channels• In well conditioned channels: MIMO (2-2) only slightly better than
Alamouti (2-2)
MIMO (2,2)
Alamouti (2,2)
• 2 TX, 2 RX• 2 bits/subcarrier• MIMO = 2 streams of
rate ½ QPSK• Alamouti = 1 stream
of rate ½ 16-QAM• ML Receivers
15 of 16March 2003 IEEE 802.20 MBWA
-2 -1 0 1 2 3 4 5 6 7 810
-3
10-2
10-1
100
Coded Eb/No (dB)
Dec
oded
Fra
me
Erro
r Rat
e
Decoded FER: Rate-1/2 Turbo-Coded QP S K, 30 mph ≤ Velocity < ∞ mph
S ingle Tx (1-1)Alamouti (2-1)TXAA-(no delay) (2-1)TXAA-(1ms ec) (2-1)TXAA-(2ms ec) (2-1)
-4 -2 0 2 4 6 810
-3
10-2
10-1
100
Coded Eb/No (dB)
Dec
oded
Fra
me
Erro
r Rat
e
Decoded FER: Rate-1/2 Turbo-Coded QP S K
S ingle Tx (1-1)Alamouti (2-1)TXAA (2-1)S ingle Tx (1-2)Alamouti (2-2)TXAA-(2-2)
Evaluating Downlink TX Diversity & TX Adaptive Arrays in Measured ChannelsEvaluating Downlink TX Diversity & TX Evaluating Downlink TX Diversity & TX Adaptive Arrays in Measured ChannelsAdaptive Arrays in Measured Channels
• 1-TX vs. 2-TX (Alamouti & TXAA)• TXAA with ideal channel knowledge• Additional Rx antenna better than additional
Tx antennas– More Rx diversity than Tx diversity in this
environment
• Effect of feedback latency on TXAA• Feedback latency of 2msec causes TXAA to
provide no gain over Alamouti for velocities > 30mph
• TXAA appropriate for stationary receivers, not appropriate for high velocity users
TXAA (2-2)Alam
outi (2-2)
(1-1)
Alamouti (2-1)
(1-2)TXAA (2-1)
(1-1)
Alamouti (2-1)
TXAA (2-1)
Increasing Feedback Delay
• Simulated Turbo Coded 2x2 OFDM using measured channels
16 of 16March 2003 IEEE 802.20 MBWA
ConclusionConclusionConclusion
• Provided a basis for future discussion & proposals on antenna array technologies for MBWA
• Antenna Array technology widely viewed as critical for future mobile broadband communication systems
– Many configurations to choose from, each with pros & cons
• Results from one set of suburban 2x2 field experiments:– Multiple receive antennas provide largest benefits– MIMO advantageous at high SNRs, high data rates, good spatial conditioning– Alamouti outperforms 2x2 MIMO at low data rates, low SNRs, low scattering– TXAA advantageous for low-mobility / portable subscribers
• Evaluating Antenna Array Technology:– Performance tends to be environmentally dependent– Realistic channel models are needed in evaluations– Evaluations should involve coded performance– Evaluations should examine system-level gains