MIMO Wireless Communication

Report
MIMO Wireless Communication
Per Hjalmar Lehne, Telenor
Guest lecture at UniK
1 March 2012
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heurestic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
2
01.03.2012
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heurestic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
3
01.03.2012
What is MIMO?
MIMO: Multiple input – multiple output
Given an arbitrary wireless communication system:
• ”A link for which the transmitting end as well as the receiving end is
equipped with multiple antenna elements”
The signals on the transmit antennas and receive antennas are ”combined” to
improve the quality of the communication (ber and/or bps)
MIMO systems use space-time processing techniques
• Time dimension is completed with the spatial dimension
4
01.03.2012
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heurestic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
5
01.03.2012
Different gains of multiple antenna systems
”Smart antenna” gain
• Beamforming to increase the average signal-to-noise (SNR)
ratio through focussing energy into desired directions
Spatial diversity gain
• Receiving on multiple antenna elements reduces fading
problems. The diversity order is defined by the number of
decorrelated spatial branches
Spatial multiplexing gain
• A matrix channel is created, opening up the possibility of
transmitting over several spatial modes of the matrix channel
increasing the link throughput at no additional frequency,
timer or power expenditure
6
01.03.2012
Multiple antenna fundamentals
Recovered data stream
Data
Tx antenna ports
Channel
Data
Rx antenna ports
Data stream
7
01.03.2012
Multiple antenna fundamentals
Recovered data stream
Data
Tx antenna ports
Data
N transmit antennas
Data stream
 h11
H  h21
 h31
h12
h13
h22
h32
h23
h33
h14 
h24 
h34 
Channel matrix
8
01.03.2012
Rx antenna ports
M receive
antennas
Multiple antenna fundamentals
Recovered data stream
Data
Tx antenna ports
A1
A2
A3
A4
Data
Rx antenna ports
Data stream
9
01.03.2012
Multiple antenna fundamentals
Spatial multiplexing
Recovered data stream
Data
Tx antenna ports
Data
Rx antenna ports
Data stream
The different data
streams are
divided in space
 h11
H  h21
 h31
h12
h13
h22
h32
h23
h33
h14 
h24 
h34 
rank(H) determines how many streams are possible to transmit
10
01.03.2012
Multiple antenna fundamentals
Transmit diversity
Recovered data stream
Data
Tx antenna ports
A1
A2
A3
A4
Data
Rx antenna ports
Data stream
Redundancy:
The data streams
contain the same
data
11
01.03.2012
Multiple antenna fundamentals
Beamforming
Recovered data stream
Data
Tx antenna ports
A1
A2
A3
A4
Data
Rx antenna ports
Data stream
12
01.03.2012
Only the best spatial
channel is used to
maximize C/N
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heurestic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
13
01.03.2012
Fundamental limits of wireless transmission
Shannon capacity of Wireless Channels:
C  log2 (1   )
• h is the unit power complex Gaussian
amplitude of the channel
C  log2 (1   h )
2
– h is a random variable
• Multiple antennas at one end:
• Capacity of MIMO Links:
Average capacity Ca
Outage capacity Co
14
01.03.2012
C  log2 (1  hh* )
 


C  log2 det I M  HH* 
N

 
PC  Co   99.9..9%
Shannon capacity of Wireless Channels
Ideal Rayleigh Channel
C  log 2 (1   h )
2
 


C  log2 det I M  HH* 
N

 
C  log2 (1  hh* )
15
01.03.2012
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heurestic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
16
01.03.2012
Data transmission over MIMO systems
Two main categories:
• Data rate maximization
– Sending as many independent signals as antennas
– Spatial multiplexing
• Diversity maximization
– The individual streams can be encoded jointly
– Protect against transmission errors caused by channel fading
– Minimize the outage probability
17
01.03.2012
Maximizing diversity with space-time block
codes
Alamouti’s scheme:
•
•
The block of symbols s0 and s1 is coded across time and
space
1
C
2
Normalization factor ensures total energy to be the same
the case of one transmitter
Reception:
•
The receiver collects the observation, y, over two symbol
periods
s  s0
s ,
0
s
*
1

s1 
T
h  h0
1  h0
ˆ
H
 *
2  h1
h1 
n
Tx0
h0
Rx
s , s 
1
18
 s0

 s1
*
0
01.03.2012
Tx1
h1
h1 
 h0* 
y0
y1   n  h  C  n
y
T
ˆ s  n
y1*  n  H
0

 s1* 
* 
s0 
Spatial multiplexing
Y  HC  N
Extending the SpaceTime Block Coding
• Transmitting
independent data over
different antennas
• The receiver must unmix the channel
• Limited diversity
benefit
19
01.03.2012
C
H
Y
Spatial multiplexing - decoding
Zero-forcing (ZF)
• Inverting matrix H
• Simple approach
• Dependent on low-correlation in H
Y  HC  N
ˆ  H 1 Y
C
Maximum likelihood (ML)
• Optimum
• Comparing all possible combination with the
observation
• High complexity
Nulling and cancelling
• Matrix inversion in layers
• Estimates one symbol, subtracts and continues
decoding successively
20
01.03.2012
ˆ  arg min Y  HC
ˆ
C
ˆ
C
Transmission scheme performance
Same transmission
rate
• Alamouti
• Spatial multiplexing –
zero forcing
• Spatial multiplexing –
maximum likelihood
• Combined STBC
spatial multiplexing
21
01.03.2012
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heuristic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
22
01.03.2012
Wireless channel modelling
The promise of high MIMO capacities largely relies on the
decorrelation properties:
• Between antennas
• Full-rankness of the MIMO channel matrix H
– E.g. spatial multiplexing becomes completely inefficient if the channel
has rank 1
Aim of channel modelling:
• Get an understanding of what performance can be reasonably
expected form MIMO systems
• To provide the necessary tools to analyze the impact of selected
antenna or propagation parameters
– Spacing, frequency, antenna height..
• To influence the system design in the best way
23
01.03.2012
Wireless channel modelling
Four approaches
• Theoretical Models
– E.g. the ”idealistic” channel matrix of perfectly uncorrelated (i.i.d.)
random Gaussian elements
• Heurestic Models
– In practice, MIMO channels will not fall completely into any of the
theoretical cases
• Broadband Channels
– Frequency selective fading is experienced a new MIMO matrix is
obtained at each frequency/sub-band
• Measured Channels
– Validate the models, provide acceptance of MIMO systems into
wireless standards
24
01.03.2012
Theoretical channel models
Ideal channel (i.i.d.):
• Corresponds to a rich multipath environment
Emphasizing the separate roles
• Antenna correlation (transmit or receive)
• Rank of the channel
– Uncorrelated High Rank (UHR aka i.i.d.)
– Correlated Low Rank (CLR)
– Antennas are placed too close to each other,
or
H  grx gtx* urx u*tx
– Too little angular spread at both transmitter
and receiver
– Uncorrelated Low Rank (ULR)
– ”pin-hole” model
25
01.03.2012
H  g rx g*tx
Heuristic channel models
Display a wide range of
MIMO channel behaviours
through the use of as few
relevant channel parameters
as possible, with as much
realism as possible
• What is the typical capacity of
a MIMO channel?
• What are the key parameters
governing capacity?
• Under what simple conditions
do we get full rank channel?
The model parameters
should be controllable or
measurable
26
01.03.2012
Antenna correlation at transmitter or receiver
A MIMO channel with correlated receive antennas:
H  R1/r 2,dr H0
• For ”large” values of the angle spread and/or
antenna spacing, R will converge to the identity
matrix
• For ”small” values of θr, dr, R becomes rank
deficient (eventually rank one) causing fully
correlated fading
Generalized model includes correlation on both
sides:
27
01.03.2012
H  R1/r2,dr H0R1/t 2,dt
The double scattering model – ”pinhole”
channels
Uncorrelated low rank:
• Significant local scattering around both the BTS and the subscriber’s
antennas
• Local scatterer’s are considered as virtual receive antennas
– When the virtual aperture is small, either on transmit or receive, the rank of the
overall MIMO channel will fall
28
01.03.2012
Broadband channels
Frequency selective channels are
experienced
MIMO capacity benefits OFDM
systems with MIMO
• Additional paths contribute to the
selectivity as well as a greater overall
angular spread
• Improving the average rank of the
MIMO channel across frequencies
H(f)
29
01.03.2012
Measured channels
Channel matrix is measured using multiple antennas at transmitter and receiver
• Results confirm the high level of MIMO capacity potential, at least in urban and
suburban areas
• Eigenvalue analysis
–
SISO
A large number of the modes of MIMO channels can be exploited to transmit data
4x4
P Kvadraturen 01 15 21
MIMO
SNR mean value and difference
800
NLOS
20
dB
LOS
10
0
600
400
0
1
2x2
MIMO
200
0
200
400
600
Route sample no.
RX= 10,14,12,16 TX= 2,6,1,5
10
bility
30
ity < C-sum)
0
Capacity Mbits/s
30
0
200
400
Route sample no.
Diversity gain, full CSI
600
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heurestic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
31
01.03.2012
System level issues – optimum use of
multiple antennas
Multiple antenna usage is not new in mobile systems:
• Spatial diversity systems
Different goals:
• Beamforming is optimum using a large number of closely spaced
antennas:
– Directional beamforming imposes stringent limits on spacing, typically a
half wavelength
– Best performance in line-of-sight (LOS)
• MIMO algorithms focusses on diversity or data rate maximization:
– Antennas will use as much space as possible to realize decorrelation
between antennas
– Turning rich multipath into an advantage and lose the gain in LOS cases
32
01.03.2012
MIMO in mobile broadband
A unfavourable aspect:
•
Increased cost and size of the subscriber’s equipment
•
Limits applicability on simple mobile devices
A better opportunity:
•
33
Wireless LAN modems – tablets - laptops
01.03.2012
Agenda
What is MIMO?
Different gains of multiple antenna systems
Fundamental Limits of Wireless Transmission
•
Shannon capacity of Wireless Channels
•
Multiple antennas at one end
•
Capacity of MIMO Links
Data transmission over MIMO Systems
•
General principles
•
Diversity using Space Time Block Codes
•
Spatial Multiplexing
Wireless channel modelling
•
Theoretical Models
•
Heurestic Models
•
Broadband Channels
•
Measured Channels
System Level Issues
•
Optimum use of multiple antennas
•
MIMO in Mobile Broadband
MIMO Transmission Scheme for HSPA and LTE
34
01.03.2012
MIMO transmission schemes for LTE
LTE supports downlink
transmissions on one, two or four
cell-specific antenna ports
• Up to two transport blocks can be
transmitted simultaneously on up to
four layers
The use of multiple antennas in
the DL of LTE comprises several
modes
The system adaptively switches
between each mode to obtain the
best possible performance as the
propagation conditions vary
35
01.03.2012
LTE Transmission modes
1
Single eNB antenna
2
Tx diversity (SFBC)
3
Open-loop SM
4
Closed-loop SM
5
Multi-user MIMO
6
Beamforming
7
UE specific RS
Downlink multi-antenna support in LTE
Up to 4x4 antennas on downlink
1
Single eNB antenna
• 8x8 on LTE-advanced
2
Tx diversity (SFBC)
3
Open-loop SM
4
Closed-loop SM
5
Multi-user MIMO
Multi-user MIMO (5)
6
Beamforming
A common physical layer architecture:
7
UE specific RS
Single-user schemes
• Transmit diversity (2)
• Spatial multiplexing (3, 4)
• Beamforming (6)
layers
code words
Scrambling
Modulation
mapper
Layer
mapper
Scrambling
36
07 September 2011
antenna ports
Modulation
mapper
Resource element
mapper
OFDM signal
generation
Resource element
mapper
OFDM signal
generation
Precoding
Downlink multi-antenna support in LTE
Up to 4x4 antennas on downlink
1
Single eNB antenna
• 8x8 on LTE-advanced
2
Tx diversity (SFBC)
3
Open-loop SM
4
Closed-loop SM
5
Multi-user MIMO
Multi-user MIMO (5)
6
Beamforming
A common physical layer architecture:
7
UE specific RS
Single-user schemes
• Transmit diversity (2)
• Spatial multiplexing (3, 4)
• Beamforming (6)
layers
code words
Scrambling
Modulation
mapper
Layer
mapper
Scrambling
37
07 September 2011
antenna ports
Modulation
mapper
Resource element
mapper
OFDM signal
generation
Resource element
mapper
OFDM signal
generation
Precoding
Transmit Diversity with 2 Tx antennas
Alamouti scheme
• Transmitted diversity streams are orthogonal:
Subcarrier (frequency)
Port (antenna)
 y 0 (1) y 0 (2)  x1 x2 
 1
   * *
1
 y (1) y (2)   x2 x1 
x1
x2
Antenna port 0
-x2*
x1*
Antenna port 1
OFDM subcarriers
38
07 September 2011
Downlink multi-antenna support in LTE
Up to 4x4 antennas on downlink
1
Single eNB antenna
• 8x8 on LTE-advanced
2
Tx diversity (SFBC)
3
Open-loop SM
4
Closed-loop SM
5
Multi-user MIMO
Multi-user MIMO (5)
6
Beamforming
A common physical layer architecture:
7
UE specific RS
Single-user schemes
• Transmit diversity (2)
• Spatial multiplexing (3, 4)
• Beamforming (6)
layers
code words
Scrambling
Modulation
mapper
Layer
mapper
Scrambling
39
07 September 2011
antenna ports
Modulation
mapper
Resource element
mapper
OFDM signal
generation
Resource element
mapper
OFDM signal
generation
Precoding
Downlink spatial multiplexing for 2x2
antennas
The number of codewords equals the transmission rank and codeword n is
mapped to layer n
Rank one precoders are column subsets of the rank two precoders
1 0 1 1  1 1 
0 1, 1  1,  j  j 

 
 

Recommendations on transmission rank and which precoder matrix to use is
obtained via feedback from the subscriber equipment (UE)
• The base station (eNB) can override the rank recommended by the UE
Codeword to layer mapping:
Codeword 1
40
Codeword 2
Rank 1
Layer 1
Rank 2
Layer 1
Layer 2
Rank 3
Layer 1
Layer 2 and 3
Rank 4
Layer 1 and 2
Layer 3 and 4
07 September 2011
Downlink multi-antenna support in LTE
Up to 4x4 antennas on downlink
1
Single eNB antenna
• 8x8 on LTE-advanced
2
Tx diversity (SFBC)
3
Open-loop SM
4
Closed-loop SM
5
Multi-user MIMO
Multi-user MIMO (5)
6
Beamforming
A common physical layer architecture:
7
UE specific RS
Single-user schemes
• Transmit diversity (2)
• Spatial multiplexing (3, 4)
• Beamforming (6)
layers
code words
Scrambling
Modulation
mapper
Layer
mapper
Scrambling
41
07 September 2011
antenna ports
Modulation
mapper
Resource element
mapper
OFDM signal
generation
Resource element
mapper
OFDM signal
generation
Precoding
DL peak throughputs in LTE
64QAM Modulation
Data rate (gross)
326Mbps
4 layer
2 layer
1 layer
42
07 September 2011
Peak Throughput
MIMO config
245Mbps
163Mbps
172.8Mbps
82Mbps
23Mbps
49Mbps
129.6Mbps
86.4Mbps
10.4Mbps
25.9Mbps
43.2Mbps
86.4Mbps
64.8Mbps
5.2Mbps
13Mbps
1.4
3
21.6Mbps
43.2Mbps
5
10
15
Carrier Bandwidth (MHz)
20
Downlink MIMO for HSPA (3G)
HSPA supports downlink closed-loop MIMO rank 2
43
07 September 2011
Other multiple antenna schemes
Multi-user (MU-) MIMO
• Spatial multiplexing to different UEs in the same cell
• Also called Spatial Division Multiple Access (SDMA)
44
07 September 2011
Summary
MIMO is using multiple antennas at both transmitter and receiver ends to set up a
wireless link
MIMO gains can be beamforming, diversity or spatial multiplexing
Wireless link capacity can be multiplied by min(M,N)
Data transmission exploits the spatial dimension by maximizing either data rate or
diversity
Wireless channel modelling is a tool to get the necessary understanding of perfoemence
and be atool to analyze the impact of the design
Optimum use of multiple antennas contain conflicting goals in the system design,
especially when it comes to antenna sizes and design
Both HSPA and LTE enables practical use of MIMO
45
01.03.2012
Literature
David Gesbert and Jabran Akhtar: ”Breaking the Barriers of Shannon’s
Capacity: An Overview of MIMO Wireless Systems”. Telektronikk, 98(1),
p53-54, 2002.
3G Americas White paper: "MIMO Transmission Schemes for LTE and
HSPA Networks”, chapter 4, p19-30. 2009.
-Extra reading for those interested:
David Gesbert etal.:” From Theory to Practice: An Overview of MIMO
Space-Time Coded Wireless Systems”. IEEE Journal on Selected Areas
in Comunications, 21(3), p281-302, April 2003.
A. Sibille, C. Oestges, A Zanella. ”MIMO: From Theory to
implementation”. Academic Press, 2010. ISBN-10: 0123821940, ISBN13: 978-0123821942
46
01.03.2012

similar documents