MATLAB Simulation Environment For Linear Modulation Communication Systems

MATLAB Simulation Environment for
Linear Modulation Communication

Systems
Pavel Loskot
loskot@ee.oulu.fi
Centre for Wireless Communications

University of Oulu Finland
11th November 2000
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
1(19)
of
O
General Features
current version: 1.0 released on 10/25/2000
general simulation environment for multiuser system
over MIMO channel
complex envelope signal representation
basic framework (I/O, data sources, signal buffers, MIMO
channel model) plus model of Tx/Rx
block-like structure
suitable for any linear modulation transceiver
(nonlinear modulation how to handle signal buffers ?)
fully parametrized
time-domain realization (universality, speed, nonlinear
techniques)
sample-wise modelling
Support
MIMO channels in sense of multiple Tx/Rx antennas
single-carrier/multi-carrier modulation w/wo direct-sequence
spreading
speed and memory optimization through simulation
scheduling
adaptive modulation variable constellation and power
MATLAB C
simulation and development of communication systems
bit/(sub)symbol/frame error rate measurements
block-wise measurements for slowly fading channels
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
2(19)
of
O
Implemented
Multicarrier Transceivers (Transmultiplexers)
1. basic OFDM (OFDM)
2. adaptive version of basic OFDM (AOFDM)
3. Cosine Modulated Filter Bank
(CMFB1 and CMFB2)
Modulation format
1. M ary P SK , M ary ASK
M ary SquareQAM , QP SK
16/64 StarQAM , 16 CircularQAM
2. adaptive version of those
3. differential modulation with differential encoding
across time or frequency:
M ary DP SK , DQP SK , 4 DQP SK ,
16/64 DAP SK
4. offset modulation - offset of imaginary part
optionally Gray encoding
pulse shaping - RRC , RC , N RZ
direct-sequence spreading (e.g. MC-CDMA)
optionally pilot symbols
continuous pilots
scattered pilots
N.B. offset modulation and filter-banks require equalization
(so far only simple channel inversion at the receiver
implemented)
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
3(19)
of
O
Implemented (2)
Constellation for 16qam
Constellation for 16dapsk/16stqam
4
2
2
0
1
2
4
4
Constellation for 16crqam
Constellation for 64dapsk/64stqam
4
3
2
1
0
0
1
2
3
4
4
Example of QAM constellations.

Biterror Rate of BPSK, QPSK, 16QAM, 64QAM over AWGN
10
simulation
10
BER
10
10
10
10
10
10
12
14
16
18
20
22
ES/N0
BER results over AWGN.

UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
4(19)
of
O
Implemented (3)
Continuous Pilot Symbols
17
16
15
14
13
12
11
10
9
8
7
6
5
PilFpos
4
3
2
1
0
10 11 12 13 14 15 16 17 18 19 20 21 22 23
Scattered Pilot Symbols
17
PilTpos
16
15
14
13
12
11
PilFpos1
10
9
8
7
6
PilFpos
5
4
3
2
1
0
2
3 4
frequency
10 11 12 13 14 15 16 17 18 19 20 21 22 23
PilTpos
PilTpos1
time
Pilot Symbols Location.

UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
5(19)
of
O
Multicarrier Modulation
Some of the transmultiplexer front ends.
Multicarrier communication system.

UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
6(19)
of
O
Channel Filtering
Signal flow.
inner MATLAB routine filter

filtering on symbol by symbol basis (universality)
maximum delay spread limited to one symbol interval
channel can be kept constant over arbitrary number of
samples
block-wise simulation for slowly fading channels

(i.e. channel generator is independently reinitialized for
each block of the simulation)
SISO, SIMO, MISO, MIMO structures available (faster)
fading can be randomly initialized but kept constant
throughout the simulation run
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
7(19)
of
O
Error Measurements
Simulation Structure
"inputsX.m"
"outputsX.m"
Run#X
SNR1: BER1
round#1
block#1
SNRK: BERK
round#X
block#X
frame#1
symbol#1
frame#X
symbol#X
subsymbol#1
subsymbol#X
The structure of simulation for measurements.
Monte Carlo measurements

a signle run - bit/(sub)symbol/frame error rate at
predefined SNR points
a single round - sets constant SNR
blocks for slowly fading channels
frames to accomodate signalling from upper layers
a symbol basic step of the simulation measurements
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
8(19)
of
O
Spreading
Nf g
Direct Sequence Spreading
Nf g
Ntg
Ntg
Nf
Tx
Nf 1
Ntg
Ntg
Nt
Nt1
DS Despreading
Nf g
Nf g
Ntg
Rx
Ntg
Nf 1
Ntg
Nf
Ntg
Nt1
Nt
Nf g
x
(de)spreading code
Ntg
(de)spreading matrix
Spreading in multicarrier modulation.
input to the transmitter still matrix

direct-sequence spreading
across time
across frequency
or both
DS-CDMA, MC-CDMA, MC-DS-CDMA
N.B. frequency hoping can be realized over subcarriers
(anyway, more or less solution in practice)
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
9(19)
of
O
System Model Structure

L-paths
data
1 pilots
1
MUX
MAP
MOD
buffer
buffer
data
Tx
MRC
Dem
awgn
DeMap
Rx
data
(1)
Nbranches pilots
MUX
MAP
MOD
buffer
buffer
NRx
NTx
Tx
MIMO
data
MRC
Dem
awgn
DeMap
(2)
Nbranches
Rx
total: NTx x NRx x L paths
awgn
No ICI
awgn
The system model and buffers.

ChBufPnt
(NRx,NTx,L,d)
Data Symbols Buffer
Output Power Buffer
(Nc,d)
(Nc,d)
buffer
buffer
w1
awgn
data
pilots
MUX
OFDM
Dem
OFDM
MOD
MAP
wN
(Nc,d)
Rx
buffer
buffer
NRx
NTx
MIMO
awgn
Tm
TxBufSym-1
Pilot Symbols Buffer
Rx
Tx
(Nc,d)
data
DeMap
RxBufSym-1
d = delay
Nc = number of subcarriers
Modulation Constellation
Buffer
The system model and buffers.
channel buffer circular

UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
10(19)
of
O
Flow Chart of the Simulation
SNR loop
while not enough errors and Xs do:
X={Bit, Subsymbol, Symbol, F rame}

init new block
PilBuf
DatBuf
SprBuf
pilots
data
spreading
symbol source
Nc
Transmitter
ModBuf
PowBuf
mapping/spreading
modulation
NT x
Fading Gener.
L NT x NRx paths
Channel
L NT x NRx
NRx
ChBuf
Receiver
demod./despread.
Ch. Estimation
Synchronization
demapping
L NT x NRx
DataEst
find XER
DataCrt
end while loop
buffer (delay)
end SNR loop
Flow chart of the simulation algorithm.
Monte Carlo measurements

in fact, buffers delays
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
11(19)
of
O
Library and Scheduling
CMFB2
CMFB1
OFDM
AOFDM
lib
alib
tx
amod
adem
rx
est
err
Open Simulation
inputsX.m
schedule
InitTx
InitMod
InitCh
InitRx
Schedulter
Execute
cmn
fch
ch
dsp
PreProcessor
Execute
Close Simulation
outputsX.m
Scheduling and executing the simulation.
facilitate stand-alone script to get

demo version (p-code)
further independent development
overcome designs from the scratch
MATLAB into C compiling
block-like structure of the final code
extendable libraries
systemlocal libraries (models of Tx and Rx)
global libraries (adaptive modulation dependend files)
common libraries (channel models, etc.)
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
12(19)
of
O
Preprocessor
Matlab Preprocessor ver. 1.0
%#ifvalid MATLAB_CONDITION
CODE1;
%#ifexist MATLAB_VARIABLE
CODE1;
%#elvalid
%#elexist
CODE2;
CODE2;
%#endvalid
%#define
%#endexist
PRP_DEFINITION
%#ifdef PRP_DEFINITION
CODE1;
%#undefine PRP_DEFINITION
%#ifundef
PRP_DEFINITION
CODE1;
%#eldef
%#elundef
CODE2;
CODE2;
%#enddef
%#endundef
Preprocessor directives.
version 1.0
text filter
i.e. conditional script filtering (not executing!)
C-like features
strong tool for simulation scheduling
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
13(19)
of
O
OFDM Transmitter/Receiver
OFDM Transmitter
cyclic
postfix
zero
padding
D/A
IFFT
data
P/S
OFDM
signal
S/P
MUX
pilots
A=f(Ps )
MAP
S/P
Nfft
Nc
cyclic
prefix
Nc = number of carriers
Nfft= FFT order
Ps = power of continuous signal at PB
Structure of OFDM transmitter.
OFDM Receiver
cyclic
postfix
from antenna 1
unused
subcarriers
A/D
w1
S/P
FFT
FEQ
detector
data
P/S
from antenna N
Nfft
cyclic
prefix
wN
Nc
channel
knowledge
timing
Nfft = FFT order

Nc = number of carriers
Structure of OFDM receiver.
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
14(19)
of
O
Filter Bank Modulation

Minus
much higher complexity than FFT-based systems
very complex design
many open problems
equalization always necessary
Plus
no guard interval needed
better spectral characteristics and efficiency
excellent flexibility
can exactly match the channel
not widely accepted by industry, yet
(Alcatel - filter-banks for DMT applications)
very general, in fact filter-banks capable to describe any
communication system scenario (special case - arbitrary
modulation in TDMA, FDMA, CDMA schemes)
increased interest into filter-banks expected in future
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
15(19)
of
O
Filter Bank Modulation (2)
Filter-bank based communication system.
Example of Cosine Modulated Filter-Bank (CMFB)

Tx filters:
fk (n) = 2hp(n) cos (2k+1)
(n )(1)k
2M
2
4
Rx filters:
N
k
hk (n) = 2hp(n) cos (2k+1)
(n )+(1)
2M
2
4
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
16(19)
of
O
On the Design of CMFB

Prototype filter
20
0
H3=3dB
20
H =21.8dB
s
40
60
70dB
80
100
120
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Fs=0.125
F3dB=0.062
Prototype filter.
j
H (ej).F (ej)
0
H1(e ).F1(e )
H (ej).F (ej)
H2(e ).F2(e )
20
40
60
80
100
120
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
CMFB frequency responses.

UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
17(19)
of
O
On the Design of CMFB (2)

Tx/Rx Filter Convolution
0.3
0.3
0.25
h1* f1
0.2
h *f
0.2
0.15
0.1
0.1
0.05
0.1
0
0.05
40
20
20
40
0.3
0.2
40
20
20
40
0.3
2
h *f
0.2
h *f
0.2
0.1
0.1
0.1
0.2
0.1
0.3
0.2
40
20
20
40
0.4
40
20
20
40
CMFB time responses.
Subchannel Crosstalk
0.3
0.1
h *f
0.25
h0* f1
0.05
0.2
0.15
0
0.1
0.05
0.05
0
0.05
40
20
20
40
0.1
40
20
40
x 10
4
h0* f2
20
20
x 10
h0* f3
4
40
20
40
6
40
20
20
40
CMFB crostalk in time domain.

UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
18(19)
of
O
Input/Output Files
SY ST EM = {OF DM, AOF DM, etc.} system
name, X simulation run
inputsX.m input file
SY ST EM coreX.m the core simulation file
SY ST EM initX.m the simulation part with initialization
SY ST EM wsX.mat saved workspace after the
initialization
outputsX.m system messages results
executable m-file to reread the results
schedf ileX.m list of files inserted into
SY ST EM coreX.m and SY ST EM initX.m
touchf ileX.m let you know the simulation is still alive
Other Features
preprocessor on/off
create stand-alone code yes/no
record the channel realization yes/no
include awg noise yes/no
Future Directions
extending of the preprocessor
(inline functions)
extending of the librariers
direct MATLAB C translation
own and yet simple translator
parameters initialization in MATLAB ,
the simulation itself in C
UNIVERSITY
FoilTEX
< >
c
Pavel
Loskot 11/11/00
19(19)
of
O

MATLAB Simulation Environment For Linear Modulation Communication Systems

Uploaded by

Copyright:

Available Formats

MATLAB Simulation Environment For Linear Modulation Communication Systems

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MATLAB Simulation Environment For Linear Modulation Communication Systems

Uploaded by

Copyright:

Available Formats

MATLAB Simulation Environment for

Linear Modulation Communication

Centre for Wireless Communications

Constellation for 16dapsk/16stqam

Constellation for 16crqam

Constellation for 64dapsk/64stqam

Example of QAM constellations.

BER results over AWGN.

Scattered Pilot Symbols

Pilot Symbols Location.

Some of the transmultiplexer front ends.

Multicarrier communication system.

inner MATLAB routine filter

block-wise simulation for slowly fading channels

The structure of simulation for measurements.

Monte Carlo measurements

Direct Sequence Spreading

Spreading in multicarrier modulation.

input to the transmitter still matrix

System Model Structure

total: NTx x NRx x L paths

The system model and buffers.

Output Power Buffer

The system model and buffers.

channel buffer circular

Flow Chart of the Simulation

while not enough errors and Xs do:

X={Bit, Subsymbol, Symbol, F rame}

end while loop

Flow chart of the simulation algorithm.

Monte Carlo measurements

Library and Scheduling

Scheduling and executing the simulation.

facilitate stand-alone script to get

Matlab Preprocessor ver. 1.0

Structure of OFDM transmitter.

Nfft = FFT order

Structure of OFDM receiver.

Filter Bank Modulation

Filter Bank Modulation (2)

Filter-bank based communication system.

Example of Cosine Modulated Filter-Bank (CMFB)

fk (n) = 2hp(n) cos (2k+1)

On the Design of CMFB

CMFB frequency responses.

On the Design of CMFB (2)

CMFB time responses.

CMFB crostalk in time domain.

You might also like