Performance Evaluation of Multi-Layered Space Frequency Time Codes - - PowerPoint PPT Presentation

performance evaluation of multi layered space frequency
SMART_READER_LITE
LIVE PREVIEW

Performance Evaluation of Multi-Layered Space Frequency Time Codes - - PowerPoint PPT Presentation

Performance Evaluation of Multi-Layered Space Frequency Time Codes for MIMO-OFDM Systems Dr. Samir Al-Ghadhban Assistant Prof. EE Dept, KFUPM, Saudi Arabia http://faculty.kfupm.edu.sa/ee/samir NCTT-MCP08 Aug, 2008 Outline Background


slide-1
SLIDE 1

Performance Evaluation of Multi-Layered Space Frequency Time Codes for MIMO-OFDM Systems

  • Dr. Samir Al-Ghadhban

Assistant Prof. EE Dept, KFUPM, Saudi Arabia

http://faculty.kfupm.edu.sa/ee/samir

NCTT-MCP08 Aug, 2008

slide-2
SLIDE 2

NCTT-MCP08 Alghadhban 2

Outline

  • Background and motivation
  • IQ-Space Frequency Time codes
  • Multi-Layered STBC vs VBLAST
  • Multi-Layered SFT Codes
slide-3
SLIDE 3

NCTT-MCP08 Alghadhban 3

Introduction: Multiple Input Multiple Output (MIMO) Channels

  • A MIMO channel is a

wireless link between MT transmit and MR receive antennas.

  • MIMO channels boost

the information capacity

  • f wireless systems by
  • rder of magnitude

[Telater95][Foschini98].

Tx Rx

Scatterers

1 2 M

T

1 2 M

R

11 1 1

( ) ( ) ( ) ( ) ( )

T R R T

M M M M

h t h t t h t h t     =       H …

slide-4
SLIDE 4

NCTT-MCP08 Alghadhban 4

Introduction: Open Loop MIMO Communication Systems

Spatial Multiplexing Transmit Diversity

D-BLAST [Fos96] V-BLAST [Wal99] Trellis Coding [Tar98] Block Coding [Ala98][Tar99a] Differential Block Coding [Tar00]

Open Loop MIMO Communication Systems

STBC G

1

STBC Combiner and Detector

B

1

1

B

  • MIMO Fading

Channel

V-BLAST Detector

B

1

B

K

1

B

  • MIMO Fading

Channel B

2

B

K-1

2

B

  • 1
K

B −

  • K

B

slide-5
SLIDE 5

NCTT-MCP08 Alghadhban 5

Multi-layered STBC is a single user system that consists of K parallel STBC

STBC G1 STBC GK SGINC Detector

B1 BK

1

B

  • K

B

  • MIMO Fading

Channel

  • It combines spatial

multiplexing with transmit diversity.

  • It is a V-BLAST

system with STBC on each layer.

slide-6
SLIDE 6

NCTT-MCP08 Alghadhban 6

How does MLSTBC compare to V-BLAST and STBC?

STBC G1 STBC GK SGINC Detector

B1 BK

1

B

  • K

B

  • MIMO Fading

Channel

V-BLAST Detector

B

1

B

K

1

B

  • MIMO Fading

Channel B

2

B

K-1

2

B

  • 1
K

B −

  • K

B

  • STBC

G

1

STBC Combiner and Detector

B

1

1

B

  • MIMO Fading

Channel

V-BLAST STBC MLSTBC

slide-7
SLIDE 7

NCTT-MCP08 Alghadhban 7

Comparison of MLSTBC and V-BLAST

  • ver 4x4 MIMO-OFDM, Nc=64 and L=4

10 20 30 40 50 10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 Es/N0 OFDM SER V-BLAST-OFDM BPSK 1/2 rate STBC-OFDM 256QAM MLSTBC-OFDM QPSK 10 20 30 40 50 10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 Es/N0 OFDM SER V-BLAST-OFDM QPSK MLSTBC-OFDM 16QAM 10 20 30 40 50 10

  • 6

10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 Es/N0 OFDM SER V-BLAST-OFDM 16QAM MLSTBC-OFDM 256QAM

slide-8
SLIDE 8

NCTT-MCP08 Alghadhban 8

Motivation

  • Pervious work on MLSTBC over MIMO-

OFDM systems didn’t take advantage of the available frequency diversity.

  • Our Goal is to design MLSTBC system that

takes full frequency diversity advantage over MIMO-OFDM channels.

  • The solution is to add space frequency time

(SFT) codes at each layer.

slide-9
SLIDE 9

NCTT-MCP08 Alghadhban 9

Design criteria of SFT codes

  • The maximum diversity available in MIMO-OFDM

systems is MTLMR [Ben Lu 2000].

  • The design criterion is to maximize the minimum

effective length and break up channel correlation in frequency domain by interleaving.

  • To achieve this diversity, the minimum effective

length of the SFT code should be equal to at least MTL, which needs large number of states for practical values.

  • For example, at MT=2 and L=3, we need 1024 states.

And at L=4, we need 16384 states

slide-10
SLIDE 10

NCTT-MCP08 Alghadhban 10

Design criteria of SFT codes

  • Our goal is to simplify the design and reduce the

number of states required to achieve the full spatial and frequency diversity.

  • Our approach is based on concatenating trellis

coded modulation (TCM) and space time block codes (STBC). [Lateif 2003]

  • Spatial diversity is guaranteed by STBC and

frequency diversity is provided by TCM.

  • We further reduce the number of states of TCM by

using IQ-TCM [AlSemari 97].

slide-11
SLIDE 11

NCTT-MCP08 Alghadhban 11

IQ-TCM [AlSemari97]

  • The minimum effective length of

TCM is upper bounded by:

min

/ 1 l v k ≤ +    

Where v is the number of memory elements and k is the number of inputs.

  • Thus, when k is reduced by

a half, lmin at most doubles and this is the reason behind the diversity increase of IQ- TCM.

I-TCM rate 1/2 Q-TCM rate 1/2

j +

In I-Dec Q-Dec

×

De

[ ]

  • 3 1 1 3

I

s ∈ −

[ ]

  • 3 1 1 3

Q

s ∈ −

{ }

1 6

  • Q

A M s∈

h h*

I

s

  • Q

s

  • 1 bps/Hz

1 bps/Hz

In: Interleaver De: De-interleaver

3

  • 1
  • 1

3 1

  • 3
  • 3

1

  • 1

3 3

  • 1
  • 3

1 1

  • 3
  • 3
  • 1

1 3 4

  • A

M S ig n a l S e t

2 bps/Hz IQ-16QAM-TCM 8-states 4AM-TCM

slide-12
SLIDE 12

NCTT-MCP08 Alghadhban 12

2 bps/Hz Comparison

I-TCM rate 1/2 Q-TCM rate 1/2

j +

In I-Dec Q-Dec

×

De

[ ]

  • 3 1 1 3

I

s ∈ −

[ ]

  • 3 1 1 3

Q

s ∈ −

{ }

1 6

  • Q

A M s∈

h h*

I

s

  • Q

s

  • 1 bps/Hz

1 bps/Hz

In: Interleaver De: De-interleaver

3

  • 1
  • 1

3 1

  • 3
  • 3

1

  • 1

3 3

  • 1
  • 3

1 1

  • 3
  • 3
  • 1

1 3 4

  • A

M S ig n a l S e t

2 bps/Hz IQ-16QAM-TCM 8-states 4AM-TCM

  • 8-states 8PSK-TCM:

v=3, k=2 lmin=2

  • 8-states IQ-16QAM-TCM:

v=3, k=1 lmin=4

min

/ 1 l v k ≤ +    

slide-13
SLIDE 13

NCTT-MCP08 Alghadhban 13

IQ-SFT

IFFT

1 2 Nc

1,1 1,2 1 1,

c

N

s s s       =        s

  • 2,1

2,2 2 2,

* * * *

Nc

s s s   −   −   − =      −     s

  • P/S

CP

IFFT

1 2

2,1 2,2 2 2,

c

N

s s s       =        s

  • 1,1

1,2 1 1,

* * * *

Nc

s s s       =          s

  • P/S

CP Time 1 Time 2

I-TCM Q-TCM

j + In

I-TCM Q-TCM

j + In

STBC

In= Interleaver CP= Cyclic Prefix P/S= Parallel to Serial converter N

c

1,I

b

1,Q

b

2,I

b

2,Q

b

1,I

s

1,Q

s

2,I

s

2,Q

s

1

s

2

s FFT 1 2 N

c

2 2 2 2

1,1 1,2 1 1, t t t t L

y y y       =        Y

  • S/P

Rem CP

Tim e 1 Tim e 2

1 1 1 1

1,1 1,2 1 1, t t t t L

y y y       =        Y

  • STBC

Com biner at each subcarrier De I-Dec Q

  • Dec

(R e) (Im )

I-Dec Q

  • Dec

(R e) (Im ) 1

s

  • 2

s

  • Encoder

Decoder

Alamouti STBC Code

slide-14
SLIDE 14

NCTT-MCP08 Alghadhban 14

Advantages of concatenated IQ-TCM-STBC at 2bps/Hz

64 4096 67108864 7 32 1024 4194304 6 16 256 262144 5 8 64 16384 4 4 16 1024 3 2 4 64 2 IQ-16QAM-STBC 8PSK-STBC Tarokh STTC QPSK L Minimum number of states to achieve full diversity (MTLMR) FCS Length

slide-15
SLIDE 15

NCTT-MCP08 Alghadhban 15

The discrete received signal over T time slots at the ith subcarrier is

1, 2, 1, 2, , , i i i i i i i i K i i K i

= +         = +         Y H S V S S H H H V S

  • Sk,i is the kth STBC at the ith

layer. Hk,iis the MR×NG MIMO matrix from group k to the receiver at the ith subcarrier.

MR: total number of receive antennas NG: number of transmit antennas per group MT: total number of transmit antennas

IQ

  • SFT

G

1

M ulti-Layered IQ

  • SFT

D ecoder IQ

  • SFT

G

K

O FD M D em

  • dulator

O FD M D em

  • dulator
slide-16
SLIDE 16

NCTT-MCP08 Alghadhban 16

Due to the short code length of STBC, the received signals over T slots are rearranged into a vector

STBC G1 STBC GK SGINC Detector

B1 BK

1

B

  • K

B

  • MIMO Fading

Channel

xk is the symbols of the kth layer. is the MIMO matrix from group k to the receiver.

1 2 1 2

ˆ ˆ ˆ ˆ

K K

= +         = +         y Hx η x x H H H η x

  • ˆ

k

H

G

M T N ⋅ ×

slide-17
SLIDE 17

NCTT-MCP08 Alghadhban 17

Serial Group Interference Nulling and Cancellation (SGINC)

  • Group interference nulling: Based on an ordering criterion,

assume that the first detected group is the kth group. Then, the algorithm calculates the orthonormal bases of the null space of:

1 1 1

ˆ ˆ ˆ ˆ

k k k K − +

  =   H H H H

  • H
  • Denote the orthonormal bases of the null space of by ,

then the received signal for the ith group after nulling is:

k

H

k

N

k k k k k

= = + y y H x η

  • N

Where is the post-processing channel matrix.

k

H

slide-18
SLIDE 18

NCTT-MCP08 Alghadhban 18

SGINC

  • STBC Combiner:
  • IQ-SFT Decoder
  • Group interference cancellation: After Decoding the kth

Layer, its contribution is subtracted from the received signal and the processing is repeated serially for each group.

  • Ordering:

– MaxMin FN – MaxAverage FN – Blind power allocation

  • Number of receive antennas should be greater than or

equal to number of layers.

H k k k

= x H y

slide-19
SLIDE 19

NCTT-MCP08 Alghadhban 19

Serial Interference cancellation/ decoding algorithm

8 9 10 11 12 13 14 15 16 17 18 10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 K= 4, MR= 4, L= 4 paths, Nc=64 and W= 4 at 8bps/Hz SNR BER Hard nulling No ordering MaxAverage Pre-FN MaxMin Pre-FN MaxAverage Post-FN MaxMin Post-FN Blind Power allocation

2dB 2dB

8 9 10 11 12 13 14 15 16 17 18 10

  • 7

10

  • 6

10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 K= 4, MR= 4, L= 4 paths, Nc=64 and W= 4 at 8bps/Hz Es/N0 BER 0 iteration 1 2 3 4 5 Perfect Cancellation

Parallel Interference Cancellation/ Decoding Algorithm

3dB 6dB

slide-20
SLIDE 20

NCTT-MCP08 Alghadhban 20

Comparison

8 9 10 11 12 13 14 15 16 17 18 10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

10 K= 4, MR= 4, L= 4 paths, Nc=64 and W= 4 at 8bps/Hz Es/N0 BER Uncoded MLSTBC-OFDM Serial: power allocation Parallel: four iterations

3dB 1dB

FD = 4 No FD

slide-21
SLIDE 21

NCTT-MCP08 Alghadhban 21

Conclusion

  • Multi-layered Space frequency time codes

were designed and evaluated over MIMO- OFDM channels.

  • The code design is simplified with IQ-TCM.
  • Serial and parallel algorithms were proposed

and evaluated for MIMO-OFDM systems.