Overview of MIMO systems

S-72. 333 Postgraduate Course in Radio Communications

Sylvain Ranvier
[email protected]

SMARAD / Radio Laboratory 1

Outline
1 Presentation
1.1 What is MIMO
1.2 Wireless channels limitations
1.3 MIMO Benefits

2 SISO Vs MIMO

3 Spatial multiplexing
3.1 Principle
3.2 Impact of channel model
3.3 V-BLAST / D-BLAST

4 Receiver design
4.1 Linear receivers for BLAST (Zero-Forcing, MMSE)
4.2 Non linear receiver (ML, SIC)
4.3 performance comparison

5 Space-Time Coding (Transmit / Receive Diversity)

6 Conclusion

SMARAD / Radio Laboratory 2

1 Presentation

Multiple-Input Multiple-Output (MIMO) Wireless Systems

1.1 What are MIMO systems ?

• A MIMO system consists of several antenna elements, plus adaptive signal
processing, at both transmitter and receiver
• First introduced at Stanford University (1994) and Lucent (1996)
• Exploit multipath instead of mitigating it

SMARAD / Radio Laboratory 3

1 Presentation

1.2 Wireless channels limitations

Wireless transmission introduces:

Fading: multiple paths with different phases add up at the receiver,

giving a random (Rayleigh/Ricean) amplitude signal.

ISI:multiple paths come with various delays, causing intersymbol

interference.

CCI: Co-channel users create interference to the target user

Noise: electronics suffer from thermal noise, limiting the SNR.

SMARAD / Radio Laboratory 4

1 Presentation

Wireless channels limitations : summary

SMARAD / Radio Laboratory 5

1 Presentation

1.3 MIMO Benefits :

• higher capacity (bits/s/Hz)

(spectrum is expensive; number of base stations limited)

• better transmission quality (BER, outage)

• Increased coverage
• Improved user position estimation

Due to :

Spatial multiplexing gain : Capacity gain at no additional power
or bandwidth consumption obtained through the use of multiple
antennas at both sides of a wireless radio link

Diversity gain : Improvement in link reliability obtained by

transmitting the same data on independently fading branches

Array gain

Interference reduction
SMARAD / Radio Laboratory 6
1 Presentation

Array gain principle :

The array gain is defined by the gain in mean SNR

SN R =
(E h ²+...+E h ²)σ ² =N E h ²σ ² =NSNR
1 N s s

σ n² σ n²
input

The output SNR is N times the input SNR

SMARAD / Radio Laboratory 7

1 Presentation

Receiving data over N antennas :

SMARAD / Radio Laboratory 8

 2 SISO Vs MIMO

2 SISO Vs MIMO

Capacity of SISO Systems (1 by 1)

At fixed time t, the SISO channel is an additive white Gaussian noise (AWGN)
channel with capacity :

C(t) = log2(1 + SNRsiso(t)) Bit/Sec/Hz

where SNRsiso(t) is the received signal to noise ratio at time t :

h(t) ² σ s ²
SNRsiso(t) =
σn ²
+3dB of extra power needed for one extra bit per transmission !

SMARAD / Radio Laboratory 9

2 SISO Vs MIMO

Capacity of MIMO systems

Note: we assume channel unknown at transmitter

 ρ ∗
 
N ΗΗ

Cerg = ε H  log2 det  IM +    ≈α min ( M, N )
    

where H is the M X N random channel matrix and ρ is the average

signal-to-noise ratio (SNR) at each receiver branch.

Capacity proportional to min of # TX and # RX antennas!

SMARAD / Radio Laboratory 10

 2 SISO Vs MIMO

Comparison : Average capacity of ideal MIMO systems

SMARAD / Radio Laboratory 11

3 Spatial multiplexing

3 Spatial multiplexing
3.1 Principle
We send multiple signals, the receiver learns the channel matrix and inverts it to
separate the data.

SMARAD / Radio Laboratory 12

 3 Spatial multiplexing

Example for 3x3:

 x1  h11 h12 h13  b1

x 2 = h 21 h 22h 23 b 2 + Noise
 x 3 h 31 h 32 h 33  b 3

bˆ1 
bˆ 2 = −1 x 2
x1
 ˆ  Η  x 3
b 3  

SMARAD / Radio Laboratory 13

3 Spatial multiplexing

3.2 Impact of channel model

MIMO Performance is very sensitive to channel matrix invertibility.

The following degrades the conditioning of the channel matrix:

Antenna correlation caused by:
- Small antenna spacing, or
- Small angle spread

Line of sight component compared with multipath fading component :

- Multipath fading component, close to random identical independent
distribution, is well conditioned
- Line of sight component is very poorly conditioned.

SMARAD / Radio Laboratory 14

 3 Spatial multiplexing

MIMO spatial multiplexing in Line-of-sight

 111 
The system Η ≈ α  111  is near rank one (non invertible) !!
 111 

Spatial multiplexing requires multipath to work !!

SMARAD / Radio Laboratory 15

3 Spatial multiplexing

3.3 V-BLAST/ D-BLAST Algorithms

(Bell-labs LAyered Space-Time architecture)
Belong to the class of Layered Space-Time Coding

• In D-BLAST, output of coders can be applied to the transmit

antennas in turn Diagonal LST coding (D-BLAST)
• In V-BLAST, output of coders operate co-channel with synchronized
symbol timing Vertical LST coding (V-BLAST)

SMARAD / Radio Laboratory 16

 4 receiver design

4 MIMO Receiver Design

4.1 Linear receivers for BLAST (Zero-Forcing, MMSE)

Zero-Forcing receiver

 x1  h11 h12 ..  S 1 

x 2 = h 21 h 22 .. S 2 + N
 :   : : :   : 

Zero Forcing implements matrix (pseudo)-inverse (ignores noise

Sˆ =Η X
#
enhancement problems) :

∗ −1
=  Η Η  Η ∗
Η
#
Where :


SMARAD / Radio Laboratory 17

4 receiver design

MMSE receiver

The MMSE (Minimum mean square error) receiver optimizes the

following criterion:
W = argmin {E |W*x – s| ²}
We find:
Ŝ = H*(HH* + Rn)-1 x
where Rn is the noise/intf covariance.

This offers a compromise between residual interference between input

signals and noise enhancement.

SMARAD / Radio Laboratory 18

 4 receiver design

4.2 Non linear receiver (ML, SIC)

Maximum likelihood receiver:

• Optimum detection
• Exhaustive search. No iterative procedure for MIMO.
• Complexity exponential in QAM order and N.

 x1  h11 h12 ..  S 1 

x 2 = h 21 h 22 .. S 2 + N
 :   : : :   : 


Maximum Likelihood Solution: Ŝ = argmin Ix – Hsl²

where s is searched over the modulation alphabet (e.g. 4QAM, 16QAM..)

SIC : Successive Interference Canceling

SMARAD / Radio Laboratory 19

4 receiver design

4.3 Performance comparison

BLAST zero-forcing vs. V-BLAST (SIC) vs BLAST-ML (2x2)

SMARAD / Radio Laboratory 20

 4 receiver design

BLAST zero-forcing vs. V-BLAST (SIC) vs BLAST-ML (4x4)

SMARAD / Radio Laboratory 21

5 Space-Time Coding (Transmit / Receive Diversity)

5 Space-Time Coding (Transmit/Receive Diversity)

Uses Transmission diversity to combat the detrimental effects in wireless fading

channels.

Three types:

•Trellis space time codes : complex but best performance in slow fading
environment (indoors).

•Layered space time codes : easy to implement but not accurate due to error
propagation effect.

•Block space time codes : best trade-off of performance vs complexity.

SMARAD / Radio Laboratory 22

 5 Space-Time Coding (Transmit / Receive Diversity)

Comparison of Performance: 2x2 STCBC and SISO

SMARAD / Radio Laboratory 23

5 Space-Time Coding (Transmit / Receive Diversity)

Comparison of Performance: V-BLAST & STCBC in MIMO-OFDM

SMARAD / Radio Laboratory 24

 5 Space-Time Coding (Transmit / Receive Diversity)

Summary : Space-Time Coding & V-BLAST

Space-Time Coding
• Space-time codes provide spatial diversity gain without requiring channel
knowledge in the transmitter
• Space-time codes do not provide array gain (due to lack of channel knowledge
in the transmitter)
• Orthogonal space-time codes decouple the vector detection problem into scalar
detection problems -> drastically simplified algorithms

V-BLAST
• Performs well when channel estimates are good
• Degradation due to channel estimation errors is fairly high
• Successive Interference Cancellation (SIC) makes for low complexity
• Danger of error propagation that is inherent of a SIC scheme
• Inferior to STBC due to lack of diversity gain at the transmitter

SMARAD / Radio Laboratory 25

6 conclusion

6 Conclusion

MIMO extremely promising but more validation work are needed :

Algorithms:
- Unifying diversity and multiplexing approaches
- Optimum loading

Low complexity receivers

- Optimum receivers (ML) are too complex
- Simple receivers (linear) give unacceptable performance at high MIMO loading

System gain evaluation

- Real gains depend on deployment scenario
- Beamforming and MIMO needs to be compared on a system level basis

SMARAD / Radio Laboratory 26

 References

• “On the Capacity of OFDM-Based Spatial Multiplexing Systems”, Helmut

Bolcskei, David Gesbert and Arogyaswami J. Paulraj, IEEE Trans.
Communications, Oct. 2001

• “An Overview of MIMO Communications- A Key to Gigabit Wireless”, A. J.

Paulraj, D. Gore, R. U. Nabar, and H. B®olcskei

• '' From theory to practice: An overview of space-time coded MIMO wireless

systems '' D. Gesbert, M. Shafi, D. Shiu, P. Smith,, IEEE Journal on Selected
Areas on Communications (JSAC). April 2003, special issue on MIMO systems.
(Recipient of the 2004 IEEE Best Tutorial Paper Award by IEEE Comm. Society).

• “Implementation of a MIMO OFDM-Based Wireless LAN System”, Allert van

Zelst, Student Member, IEEE, and Tim C. W. Schenk, Student Member, IEEE

• “MIMO Systems With Antenna Selection”, Andreas F.Molish, Moe Z.Win

SMARAD / Radio Laboratory 27

Homework

1. Explain the principle of spatial multiplexing.

2. Describe briefly what happens in MIMO spatial multiplexing if there is

just line of sight without multipath ?

SMARAD / Radio Laboratory 28