3G 04 DigitalComm

Wireless Information Transmission System Lab.
National Sun Yat-sen University Institute of Communications Engineering

Introduction to Digital
Communications System
2
Recommended Books
Digital Communications / Fourth Edition (textbook)
-- John G. Proakis, McGraw Hill
Communication Systems / 4th Edition
-- Simon Haykin, John Wiley & Sons, Inc.
Digital Communications Fundamentals and Applications /
2nd Edition
-- Bernard Sklar, Prentice Hall
Principles of Communications / Fifth Edition
-- Rodger E. Ziemer and William H. Tranter, John Wiley &
Sons, Inc.
Modern Digital and Analog Communication Systems
-- B.P. Lathi, Holt, Rinehart and Winston, Inc.
3
Example of Communications System
Switch
Transmission
Equipment
Local
Loop
regenerator
T1/E1 Facilities
Switch
Transmission
Equipment
Local
Loop
regenerator
T1/E1 Facilities
M
U
X
SONET
SDH
Public Switched Telephone Network (PSTN)
A/D Conversion
(Digitization)
A/D Conversion
(Digitization)
A/D Conversion
(Digitization)
Central Office
Central Office
Central Office
T1/E1 Facilities
Switch
Transmission
Equipment
Local
Loop
regenerator
T1/E1 Facilities
Mobile
Switching
Center
Mobile
Switching
Center
Base
Station
Base
Station
4
Basic Digital Communication Nomenclature
Textual Message: information comprised of a sequence of
characters.
Binary Digit (Bit): the fundamental information unit for all
digital systems.
Symbol (m
i
where i=1,2,M): for transmission of the bit
stream; groups of k bits are combined to form new symbol
from a finite set of Msuch symbols; M=2
k
.
Digital Waveform: voltage or current waveform representing
a digital symbol.
Data Rate: Symbol transmission is associated with a symbol
duration T. Data rate R=k/T [bps].
Baud Rate: number of symbols transmitted per second [baud].
5
Nomenclature Examples
6
Messages, Characters, and Symbols
7
Typical Digital Communications System
Information Bits
) ( t s
i
Format
Source
Encoding
Encryption
Channel
Encoding
Multiplexing Modulation
Frequency
Spreading
Multiple
Access
TX
RF
PA
Bit
Stream
Digital
Waveform
Synchronization
C
H
A
N
N
E
L
Source Bits Channel Bits
Channel Bits Source Bits
) (t s
i
Digital
Input
i
m
Digital
Output
i
m
Information Sink
From Other Sources
To Other Destinations
Optional
Essential
Interleaving
Format
Source
Decoding
Decryption
Channel
Decoding
Demultiplexing Demodulation
Frequency
Despreading
Multiple
Access
RX
RF
IF
Deinterleaving
Format
9
I nformat ion Bit s
) ( t s
i
Format
Source
Encoding
Encrypt ion
Channel
Encoding
Mult iplexing Modulat ion
Frequency
Spreading
Mult iple
Access
TX
RF
PA
Bit
St ream
Digit al
Waveform
Synchronizat ion
C
H
A
N
N
E
L
Source Bit s Channel Bit s
Channel Bit s Source Bit s
) (t s
i
Digit al
I nput
i
m
Digit al
Out put
i
m
I nformat ion Sink
From Ot her Sources
To Ot her Dest inat ions
Opt ional
Essent ial
I nt erleaving
Format
Source
Decoding
Decrypt ion
Channel
Decoding
Demult iplexing Demodulat ion
Frequency
Despreading
Mult iple
Access
RX
RF
I F
Deint erleaving
10
Formatting and Baseband Transmission
11
Sampling Theorem
12
Sampling Theorem
Sampling Theorem: A bandlimited signal having no
spectral components above f
m
hertz can be determined
uniquely by values sampled at uniform intervals of T
s
seconds, where
In sample-and-hold operation, a switch and storage
mechanism form a sequence of samples of the
continuous input waveform. The output of the sampling
process is called pulse amplitude modulation (PAM).
m
S
f
T
2
1
m S
f f 2 rate sampling or
13
Sampling Theorem
=
= =
n
S
S
S
nf f X
T
f X f X f X ) (
1
) ( ) ( ) (

14
Spectra for Various Sampling Rates
15
Natural Sampling
16
Pulse Code Modulation (PCM)
PCM is the name given to the class of baseband
signals obtained from the quantized PAM signals by
encoding each quantized sample into a digital word.
The source information is sampled and quantized to
one of L levels; then each quantized sample is digitally
encoded into an -bit (=log
2
L) codeword.
17
Example of Constructing PCM Sequence
18
Uniform and Non-uniform Quantization
19
Statistical Distribution of Single-Talker
Speech Amplitudes
50% of the time, speech voltage is less than RMS.
Only 15% of the time, voltage exceeds RMS.
Typical voice signal dynamic range is 40 dB.
20
Problems with Linear Quantization
Fact: Unacceptable S/N for small signals.
Solution:
Increasing quantization levels price is too high.
Applying nonlinear quantization achieved by first
distorting the original signal with a logarithmic
compression characteristic and then using a uniform
quantizer.
At the receiver, an inverse compression characteristic,
called expansion, is applied so that the overall
transmission is not distorted. The processing pair is
referred to as companding.
21
Implementation of Non-linear Quantizer
22
Companding Characteristics
In North America: -law compression:
In Europe: A-law compression:
<
+
=
+
+
=
0 for 1
0 for 1
sgn
where
sgn
) 1 ( log
)] / ( 1 [ log
max
max
x
x
x
x
x x
y y
e
e
<
+
+
<
+
=
1
1
sgn
log 1
)] / ( [ log 1
1
0 sgn
log 1
) / (
max
max
max
max
max
max
x
x
A
x
A
x x A
y
A x
x
x
A
x x A
y
y
e
e
e
23
Compression Characteristics
Standard values of is 255 and A is 87.6.
Source Coding
25
Information Bits
) ( t s
i
Format
Source
Encoding
Encryption
Channel
Encoding
Frequency
Spreading
Multiple
Access
TX
RF
PA
Bit
Stream
Digital
Waveform
Synchronization
C
H
A
N
N
E
L
) (t s
i
Digital
Input
i
m
Digital
Output
i
m
Information Sink
From Other Sources
Optional
Essential
Interleaving
Format
Source
Decoding
Decryption
Channel
Decoding
Frequency
Despreading
Multiple
Access
RX
RF
IF
Deinterleaving
26
Source Coding
Source coding deals with the task of forming efficient
descriptions of information sources.
For discrete sources, the ability to form reduced data
rate descriptions is related to the information content
and the statistical correlation among the source
symbols.
For analog sources, the ability to form reduced data
rate descriptions, subject to a fixed fidelity criterion I
related to the amplitude distribution and the temporal
correlation of the source waveforms.
27
Huffman Coding
The Huffman code is source code whose average word
length approaches the fundamental limit set by the
entropy of a discrete memoryless source.
The Huffman code is optimum in the sense that no other
uniquely decodable set of code-words has smaller
average code-word length for a given discrete
memoryless source.
28
Huffman Encoding Algorithm
1. The source symbols are listed in order of decreasing
probability. The two source symbols of lowest
probability are assigned a 0 and a 1.
2. These two source symbols are regarded as being
combined into a new source symbol with probability
equal to the sum of the two original probabilities. The
probability of the new symbol is placed in the list in
accordance with its value.
3. The procedure is repeated until we are left with a final
list of source statistics of only two for which a 0 and a 1
are assigned.
4. The code for each (original) source symbol is found by
working backward and tracing the sequence of 0s and 1s
assigned to that symbol as well as its successors.
29
Example of Huffman Coding
Symbol
S0
S1
S2
S3
S4
Probability
0.4
0.2
0.2
0.1
0.1
Code Word
00
10
11
010
011
Symbol
S0
S1
S2
S3
S4
Stage 1
0.4
0.2
0.2
0.1
0.1
Stage 2
0.4
0.2
0.2
0.2
Stage 3
0.4
0.4
0.2
Stage 4
0.6
0.4
0
1
0
1
0
1
0
1
30
Properties of Huffman Code
Huffman encoding process is not unique.
Code words for different Huffman encoding process
can have different lengths. However, the average
code-word length is the same.
When a combined symbol is moved as high as
possible, the resulting Huffman code has a
significantly smaller variance than when it is moved
as low as possible.
Huffman code is a prefix code.
A prefix code is defined as a code in which no code-word
is the prefix of any other code-word.
31
Bit Compression Technologies for Voice
Differential PCM (DPCM)
Adaptive DPCM
Delta Modulation (DM)
Adaptive DM (ADM)
Speech Encoding
.
.
.
32
Differential PCM (DPCM)
33
Delta Modulation (DM)
Delta modulation is a one-bit DPCM.
Advantage: bit compression.
Disadvantage: slope overload.
34
Speech Coding Objective
Reduce the number of bits needed to be transmitted,
therefore lowering the bandwidth required.
35
Speech Properties
Voiced Sound
Arises in generation of vowels and latter portion of some consonants.
Displays long-term repetitive pattern corresponding to the duration of a
pitch interval
Pulse-like waveform.
Unvoiced Sound
Arises in pronunciation of certain consonants such as s, f, p, j,
x, , etc.
Noise-like waveform.
36
Categories of Speech Encoding
Waveform Encoding
Treats voice as analog signal and does not use properties of
speech:
Source Model Coding or Vocoding
Treats properties of speech to preserve word information
Hybrid or parametric methods
Combines waveform and vocoding
37
Linear Predictive Coder (LPC)
38
Multi-Pulse Linear Predictive Coder
(MP-LPC)
39
Regular Pulse Excited Long Term Prediction
Coder (RPE-LPT)
40
Code-Excited Linear Predictive (CELP)
41
Speech Coder Complexity
42
Speech Processing for GSM
Composition of the 13 kbps signal:
36 bits for filter parameters every 20 ms.
9 bits for LTP every 5 ms.
47 bits for RPE every 5 ms.
Thus, in a 20 ms (2080-bit block, or 260 sample) interval,
we need a total of
36+9*20/5+47*20/5=260 bits.
Data Rate = 260/(20 ms) = 13 kbps.
43
Speech Processing for IS-54
Composition of the 7.95 kbps signal:
43 bits for filter parameters every 20 ms.
7 bits for LTP every 5 ms.
88 bits for codebook every 20 ms.
Thus, in a 20 ms (2080-bit block, or 260 samples) interval, we
need a total of:
43+7*20/5+88=159 bits.
Data Rate = 159/(20ms) = 7.95 kbps.
Channel Coding
45
Information Bits
) ( t s
i
Format
Source
Encoding
Encryption
Channel
Encoding
Frequency
Spreading
Multiple
Access
TX
RF
PA
Bit
Stream
Digital
Waveform
Synchronization
C
H
A
N
N
E
L
) (t s
i
Digital
Input
i
m
Digital
Output
i
m
Information Sink
From Other Sources
Optional
Essential
Interleaving
Format
Source
Decoding
Decryption
Channel
Decoding
Frequency
Despreading
Multiple
Access
RX
RF
IF
Deinterleaving
46
Channel Coding
Error detecting coding: Capability of detecting errors so
that re-transmission or dropping can be done.
Cyclic Redundancy Code (CRC)
Error Correcting Coding: Capability of detecting and
correcting errors.
Block Codes: BCH codes, RS codes, etc.
Convolutional codes.
Turbo codes.
47
Linear Block Codes
Encoder transforms block of k successive binary digits
into longer block of n (n>k) binary digits.
Called an (n,k) code.
Redundancy = n-k; Code Rate = k/n;
There are 2
k
possible messages.
There are 2
k
possible code words corresponding to the
messages.
Code word (or code vector) is an n-tuple from the space
V
n
of all n-tuple.
Storing the 2
k
code vector in a dictionary is prohibitive
for large k.
48
Vector Spaces
The set of all binary n-tuples, V
n
, is called a vector
space over GF (2).
GF: Galois Field.
Two operations are defined:
Addition:
Scalar Multiplication:
Example: Vector Space V
4
0000 0001 0010 0011 0100 0101 0110 0111
1000 1001 1010 1011 1100 1101 1110 1111
(0101)+(1110)=(0+1, 1+1, 0+1, 1+0)=(1, 0, 1, 1)
1(1010)=(11, 10, 11, 10)=(1, 0, 1, 0)
n n
U V U V U V U V + + + + + + = + ...
2 2 1 1
n
aV aV aV V a + + + = ...
2 1
49
Subspaces
A subset S of V
n
is a subspace if
The all-zero vector is in S
The sum of any two vectors in S is also in S.
Example of S:
1111
1010
0101
0000
3
2
1
0
=
=
=
=
V
V
V
V
50
Reducing Encoding Complexity
Key feature of linear block codes: the 2
k
code vectors
form a k-dimensional subspace of all n-tuples.
Example: k = 3, 2
k
= 8, n = 6, ( 6 , 3 ) code
Message Code Word
0 0 0 0 0 0 0 0 0
1 0 0 1 1 0 1 0 0
0 1 0 0 1 1 0 1 0
1 1 0 1 0 1 1 1 0
0 0 1 1 0 1 0 0 1
1 0 1 0 1 1 1 0 1
0 1 1 1 1 0 0 1 1
1 1 1 0 0 0 1 1 1
tuples. - 6 all of space vector the
of subspace l dimensiona - 3 A
51
Reducing Encoding Complexity
1 2
1 2
It is possible to find a set of linearly independent -
tuples , , ..., such that each -tuple of the suspace
is a linear combination of , , ..., .
k
k
k n
v v v n
v v v
k i
m
v m v m v m u
i
k k
,..., 1
1 or 0 where
... word Code
2 2 1 1
=
=
+ + + =
52
Generator Matrix
The 2
k
code vectors can be described by a set of k linearly
independent code vectors.
Let m=[m
1
, m
2
, , m
k
] be a message.
Code word corresponding to message m is obtained by:
1 11 12 1
2 21 22 2
1 2
Generator Matrix
n
n
k k k kn
v v v v
v v v v
G k n
v v v v

= = =

. . . .
[ ]
= =
k
k
v
v
v
m m m G m u
.
2
1
2 1

53
Generator Matrix
Storage is greatly reduced.
The encoder needs to store the k rows of G instead of
the 2
k
code vectors of the code.
For example:
[ ]
[ ] [ ] [ ] [ ]
1
2
3
1 1 2 3
2
3
1 1 0 1 0 0
Let 0 1 1 0 1 0 and 1 1 0
1 0 1 0 0 1
Then
1 1 0
1 1 0 1 110100 1 011010 0 101001
[1 0 1 1 1 0] Code Vector for 11
v
v m
v
v v v v
u v
v m

= = =

= + +

= = + +

= =

G
[ ]
0
54
Systematic Code
55
Parity Check Matrix
For each generator matrix G, there exists a parity check matrix H
such that the rows of G are orthogonal to the rows of H. (uh=0)
U is a code word generated by matrix G if and only if uH
T
=0
1 11 12 1
2 21 22 2
( ) ( )1 ( )2 ( )
1 2
1 1 2 2
, , ,
0
where 1, 2, ,
n
n
n k n k n k n k n
n
T
i i n in
h h h h
h h h h
h h h h
u u u u
uH u h u h u h
i n k

= =

=
= + + + =
=
H
. . . .
56
Parity Check Matrix and Syndrome
In a systematic code with G=[P
kxr
I
kxk
]
H=[I
rxr
P
T
rxk
]
Received Code Error

Vector Vector Vector
u e r
= +

T
H r s
r
=
correction and detection error for used of Syndrome
=
Otherwise 0
vector code a is If 0
Syndrome
r
s
57
Example of Syndrome Test
The 6-tuple 1 0 1 1 1 0 is the code vector corresponding to the
message 1 1 0.
Compute the syndrome for the non-code-vector 0 0 1 1 1 0
[ ]
1 1 0 1 0 0
0 1 1 0 1 0
1 0 0 1 0 1
1 0 1 0 0 1
0 1 0 1 1 0
0 0 1 0 1 1
k
T
n k
=

P I
H I P
G
H
_ _
[ ] [ ]
1 0 0
0 1 0
0 0 1
1 0 1 1 1 0 0 0 0
1 1 0
0 1 1
1 0 1
T
s u

= = =

H
[ ] [ ]
T
0 0 1 1 1 0 1 0 0 s = = H
58
Weight and Distance of Binary Vectors
Hamming Weight of a Vector:
w(v) = Number of non-zero bits in the vector.
Hamming Distance between 2 vectors:
d(u,v) = Number of bits in which they differ.
For example: u=10010110001
v=11001010101
d(u,v) = 5.
d(u,v) =w(u+v)
The Hamming Distance between 2 vectors is equal to the
Hamming Weight of their vector sum.
59
Minimum Distance of a Linear Code
The set of all code vectors of a linear code form a
subspace of the n-tuple space.
If u and v are 2 code vectors, then u+v must also be a
code vector.
Therefore, the distance d(u,v) between 2 code vectors
equals the weight of a third code vector.
d(u,v) =w(u+v)=w(w)
Thus, the minimum distance of a linear code equals
the minimum weight of its code vectors.
A code with minimum distance d
min
can be shown to
correct (d
min
-1)/2 erroneous bits and detect (d
min
-1)
erroneous bits.
60
Example of Minimum Distance
d
min
=3
61
Example of Error Correction and Detection
Capability
u
v
Strength Detecting Error : 1
Strength Correcting Error :
2
1
7 ) , (
min max
min
max
min
=

=
=
d m
d
t
v u d
62
Convolutional Code Structure
1 2 k 1 2 k 1 2 k
1 2 K
+ + + +
k bits
1 2 n-1 n
Output
63
Convoltuional Code
Convolutional codes
k = number of bits shifted into the encoder at one time
k=1 is usually used!!
n = number of encoder output bits corresponding to the k
information bits
r = k/n = code rate
K = constraint length, encoder memory
Each encoded bit is a function of the present input bits
and their past ones.
64
Generator Sequence
. 1 and , 1 , 0 , 1
) 1 (
3
) 1 (
2
) 1 (
1
) 1 (
0
= = = = g g g g
Generator Sequence: g
(1)
=(1 0 1 1)
r
0
r
2
r
1
u
v
r
0
r
2
r
1
u
v
r
3
. 1 and 0, , 1 , 1 , 1
) 2 (
4
) 2 (
3
) 2 (
2
) 2 (
1
) 2 (
0
= = = = = g g g g g
Generator Sequence: g
(2)
=(1 1 1 0 1)
65
Convolutional Codes
An Example (rate=1/2 with K=2)
00 00 0
Present Next Output
00
1 00 10 11
01 0
1
0
1
0
1
01
00
10
10 01
10 11
11
11
01
11
11
00
01
10
10
01
x
1
x
2
G
1
(x)=1+x
2
G
2
(x)=1+x
1
+x
2
00
01 10
11
1(11)
0(01)
1(00)
1(01)
0(00)
0(11)
0(10) 1(10)
State Diagram
66
Trellis Diagram Representation
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Trellis termination: K tail bits with value 0 are usually added to the end of the code.
67
Encoding Process
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Input: 1 0 1 1 1 0 0
Output: 11 01 00 10 01 10 11
68
Viterbi Decoding Algorithm
Maximum Likelihood (ML) decoding rule
An efficient search algorithm
Performing ML decoding rule.
Reducing the computational complexity.
received sequence r
ML
detected sequence d
min(d,r) !!
69
Basic concept
Generate the code trellis at the decoder
The decoder penetrates through the code trellis level by level in
search for the transmitted code sequence
At each level of the trellis, the decoder computes and
compares the metrics of all the partial paths entering a node
The decoder stores the partial path with the larger metric and
eliminates all the other partial paths. The stored partial path is
called the survivor.
70
Viterbi Decoding Process
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Output: 11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
71
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Output: 11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
4
2
1
1
72
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Output: 11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
4
2
1
1
3
2
1
2
73
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Output: 11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
4
2
1
1
3
2
1
2
3
2
3
1
74
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Output: 11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
4
2
1
1
3
2
1
2
3
2
3
1
3
3
3
1
75
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Output: 11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
4
2
1
1
3
2
1
2
3
2
3
1
3
3
3
1
3
2
76
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Output: 11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
4
2
1
1
3
2
1
2
3
2
3
1
3
3
3
1
3
2
2
77
00 00
10
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
00
01
00
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
0
(
1
0
)
1(01)
1
(
1
0
)
1
(
0
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
1
(
1
1
)
1
(
1
0
)
0
(
0
1
)
0(00)
1
(
1
1
)
0(00)
0
(
1
0
)
0
(
1
1
)
0
(
0
1
)
0(00)
0
(
1
1
)
Decision:11 01 00 10 01 10 11
Receive: 11 11 00 10 01 11 11
2
0
4
2
1
1
3
2
1
2
3
2
3
1
3
3
3
1
3
2
2
78
Channel Coding in GSM
79
Channel Coding in IS-54/136
80
Turbo Codes Basic Concepts
Turbo coding uses parallel concatenation of two
recursive systematic convolutional codes joined through
an interleaver.
Information bits are encoded block by block.
Turbo codes uses iterative decoding techniques.
Soft-output decoder is necessary for iterative decoding.
Turbo codes can approach to Shannon limit.
81
Turbo Codes Encoder - An Example
When the switch is placed on the low position, the tail bits are feedback
and the trellis will be terminated.
X(t)
X(t)
Y(t)
Interleaver
X'(t)
Y(t)
82
A systematic convolutional encoder with memory 2
The dotted line is for termination code
Test sequence: 1011
D D
X
0
X
1
1 1 0 1
Turbo Codes Encoding Example
83
0 0
X
0
=1
X
1
=1
1 1 0 1
00
01
10
11
1 1
84
1 0
X
0
=0
X
1
=1
1 1 0
00
01
10
11
1 1
0 1
85
1 1
X
0
=1
X
1
=0
1 1
00
01
10
11
1 1
0 1
1 0
86
1 1
X
0
=1
X
1
=0
1
00
01
10
11
1 1
0 1
1 0 1 0
87
1 1
X
0
=0
X
1
=1
00
01
10
11
1 1
0 1
1 0 1 0
0 1
88
0 1
X
0
=1
X
1
=1
00
01
10
11
1 1
0 1
1 0 1 0
0 1
1 1
89
0 0
X
0
=0
X
1
=0
00
01
10
11
1 1
0 1
1 0 1 0
0 1
1 1
0 0
90
D D
X
0
X
1
D D
(X
0
)
X
2
Interleaver
Output sequence: X
0
, X
1
, X
2
, X
0
, X
1
, X
2
, X
0
, X
1
, X
2
,...
1 1 0 1
1 0 1 1
91
The second encoder input is the interleaved
data
0 1
1 1
1 0 1 1
1 1 0 1
00
01
10
11
1 1
1 0 0 0 1 0
1 1
0 0 0 0
92
CRC in WCDMA
g
CRC24
(D) = D
24
+ D
23
+ D
6
+ D
5
+ D + 1;
g
CRC16
(D) = D
16
+ D
12
+ D
5
+ 1;
g
CRC12
(D) = D
12
+ D
11
+ D
3
+ D
2
+ D + 1;
g
CRC8
(D) = D
8
+ D
7
+ D
4
+ D
3
+ D + 1.
93
Channel Coding Adopted in WCDMA
No coding
1/3 Turbo coding
1/3, 1/2
CPCH, DCH, DSCH,
FACH
RACH
PCH 1/2
Convolutional
coding
BCH
Coding rate Coding scheme Type of TrCH
94
Convolutional Coding in WCDMA
Output 0
G
0
= 557 (octal)
Input
D D D D D D D D
Output 1
G
1
= 663 (octal)
Output 2
G
2
= 711 (octal)
Output 0
G
0
= 561 (octal)
Input
D D D D D D D D
Output 1
G
1
= 753 (octal)
(a) Rate 1/2 convolutional coder
(b) Rate 1/3 convolutional coder
95
Turbo Coder in WCDMA
x
k
x
k
z
k
Turbo code
internal interleaver
x
k
z
k
D
D D D
D D
Input
Output
Input
Output
x
k
1st constituent encoder
2nd constituent encoder
Interleaving
97
Information Bits
) ( t s
i
Format
Source
Encoding
Encryption
Channel
Encoding
Frequency
Spreading
Multiple
Access
TX
RF
PA
Bit
Stream
Digital
Waveform
Synchronization
C
H
A
N
N
E
L
) (t s
i
Digital
Input
i
m
Digital
Output
i
m
Information Sink
From Other Sources
Optional
Essential
Interleaving
Format
Source
Decoding
Decryption
Channel
Decoding
Frequency
Despreading
Multiple
Access
RX
RF
IF
Deinterleaving
98
Bursty Error in Fading Channel
99
Interleaving Mechanism (1/2)
Bit
Interleaver
x
y
y
j x n-bit
Shift registers
x
Write Clock Read Clock
Bit Stream before entering bit interleaver:
x=(a
11
a
12
a
1n
a
21
a
22
a
2n
a
j1
a
j2
a
jn
)
100
Interleaving Mechanism (2/2)
Conceptually, the WRITE clock places the bit stream
x by the row while the REA clock takes the bit stream
y by the column:
Bit stream at the output of the bit interleaver:
jn j j
n
n
a a a
a a a
a a a
. . .
. . . . . .
. . . . . .
. . . . . .
. . .
. . .
2 1
2 22 21
1 12 11
( )
jn n n j j
a a a a a a a a a y ... ... ... ...
2 1 2 22 12 1 21 11
=
101
Burst Error Protection with Interleaver
Modulation
103
Information Bits
) ( t s
i
Format
Source
Encoding
Encryption
Channel
Encoding
Frequency
Spreading
Multiple
Access
TX
RF
PA
Bit
Stream
Digital
Waveform
Synchronization
C
H
A
N
N
E
L
) (t s
i
Digital
Input
i
m
Digital
Output
i
m
Information Sink
From Other Sources
Optional
Essential
Interleaving
Format
Source
Decoding
Decryption
Channel
Decoding
Frequency
Despreading
Multiple
Access
RX
RF
IF
Deinterleaving
104
Modulation
Digital Modulation: digital symbols are transformed into
waveforms that are compatible with the characteristics of the
channel.
In baseband modulation, these waveforms are pulses.
In bandpass modulation, the desired information signal
modulates a sinusoid called a carrier. For radio transmission,
the carrier is converted in an electromagnetic (EM) wave.
Why modulation?
Antenna size should be comparable with wave length
baseband transmission is not possible.
Modulation may be used to separate the different signals
using a single channel.
105
PCM Waveform Representations
106
PCM Waveform Representations
PCM waveform is also called line codes.
Digital baseband signals often use line codes to provide
particular spectral characteristics of a pulse train.
NRZ-L.
NRZ-M.
NRZ-S.
Unipolar-RZ.
Polar-RZ.
Bi--L.
Bi--M.
Bi--S.
Dicode-NRZ.
Dicode-RZ.
Delay Mode.
4B3T.
Multi-level.
etc.
107
PCM Waveform : NRZ-L
NRZ Level (or NRZ Change)
One is represented by one level.
Zero is represented by the other level.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
108
PCM Waveform : NRZ-M
NRZ Mark (Differential Encoding)
One is represented by a change in level.
Zero is represented by a no change in level.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
109
PCM Waveform : NRZ-S
NRZ Space (Differential Encoding)
One is represented by a no change in level.
Zero is represented by a change in level.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
110
PCM Waveform : Unipolar-RZ
Unipolar - RZ
One is represented by a half-bit width pulse.
Zero is represented by a no pulse condition.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
111
PCM Waveform : Polar-RZ
Polar - RZ
One and Zero are represented by opposite
level polar pulses that are one half-bit in width.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
112
PCM Waveform : Bi--L
Bi--L (Biphase Level or Split Phase Manchester
11 + 180
o
)
One is represented by a 10.
Zero is represented by a 01.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
113
PCM Waveform : Bi--M
Bi--M ( Biphase Mark or Manchester 1)
A transition occurs at the beginning of every bit period.
One is represented by a second transition one half bit
period later.
Zero is represented by no second transition.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
114
PCM Waveform : Bi--S
Bi--S ( Biphase Space)
A transition occurs at the beginning of every bit period.
One is represented by no second transition.
Zero is represented by a second transition one-half bit
period later.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
115
PCM Waveform : Dicode - NRZ
Dicode Non-Return-to-Zero
A One to Zero or Zero to One changes polarity.
Otherwise, a Zero is sent.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
116
PCM Waveform : Dicode - RZ
Dicode Return-to-Zero
A One to Zero or Zero to One transition produces
a half duration polarity change.
Otherwise, a Zero is sent.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
117
PCM Waveform : Delay Mode
Dicode Non-Return-to-Zero
A One is represented by a transition at the midpoint of
the bit interval.
A Zero is represented by a no transition unless it is
followed by another zero. In this case, a transition is
placed at the end of bit period of the first zero.
1 0 1 1 0 0 0 1 1 0 1
0
+E
-E
118
PCM Waveform : 4B3T
O --
119
PCM Waveform : 4B3T
Ternary words in the middle column are balanced in
their DC content.
Code words from the first and third columns are selected
alternately to maintain DC balance.
If more positive pulses than negative pulses have been
transmitted, column 1 is selected.
Notice that the all-zeros code word is not used.
120
PCM Waveform : Multilevel Transmission
Multilevel transmission with 3 bits per signal interval.
121
Criteria for Selecting PCM Waveform
DC component: eliminating the dc energy from the
signals power spectrum.
Self-Clocking: Symbol or bit synchronization is
required for any digital communication system.
Error detection: some schemes provide error detection
without introducing additional error-detection bits.
Bandwidth compression: some schemes increase
bandwidth utilization by allowing a reduction in
required bandwidth for a given data rate.
Noise immunity.
Cost and complexity.
122
Spectral Densities of Various PCM Waveforms
123
Linear Modulation Techniques
Digital modulation techniques may be broadly classified as linear
and nonlinear.
In linear modulation techniques, the amplitude of the transmitted
signal, s(t), varies linearly with the modulating digital signal, m(t).
Linear modulation techniques are bandwidth efficient, though
they must be transmitted using linear RF amplifiers which have
poor power efficiency.
Using power efficient nonlinear amplifiers leads to the
regeneration of filtered sidelobes which can cause severe adjacent
channel interference, and results in the loss of all the spectral
efficiency gained by linear modulation.
Clever ways have been developed to get around these difficulties:
QPSK, OQPSK, /4-QPSK.
124
Digital Modulations
Basic digital modulated signal:
v(t) = A(t) cos (t + )
Where A(t) = Amplitude; = Frequency; = Phase;
125
Basic Digital Modulations
126
Extended Modulated Signals M-FSK
Example: 16-FSK
Every 4 bits is encoded as: 16 , , 2 , 1 ) cos( = j t A
j
Gray Coding.
127
Extended Modulated Signals M-PSK
Example: 16-PSK
Every 4 bits is encoded as:
sin( ) 1, 2, ,16
j
A t j + =
Dotted lines are decision boundaries.
Gray Coding.
128
Extended Modulated Signals 16-QAM
Every 4 bits is represented by one point in the signal constellation.
Every point has its unique amplitude and phase.
129
Binary Phase Shift Keying (BPSK)
In BPSK, the phase of a constant amplitude carrier signal is
switched between two values according to the two possible
signals m
1
and m
2
corresponding to binary 1 and 0. Normally,
the two phases are separated by 180
o
.
( ) ( ) ( )
( ) ( ) { }
( ) ( )
( )
( )
( )
( )
( )
( )
( )
2
2 2
2
cos 2 0
Re exp 2
2 sin
2
sin sin
2
c
BPSK
b
BPSK c c b
b
BPSK c
j
b b
BPSK b
g t
b b
c b c b
b
BPSK
c b c b
E
s t m t f t t T
T
g t j f t
E fT
g t m t e P f E
T fT
f f T f f T
E
P f
f f T f f T

= +
=

= =

= +

130
Power Spectral Density (PSD) of a BPSK
Signal.
131
BPSK Receiver
BPSK uses coherent or synchronous demodulation,
which requires that information about the phase and
frequency of the carrier be available at the receiver.
If a low level pilot carrier signal is transmitted along
with the BPSK signal, then the carrier phase and
frequency may be recovered at the receiver using a
phase locked loop (PLL).
If no pilot carrier is transmitted, a Costas loop or
squaring loop may be used to synthesize the carrier
phase and frequency from the received BPSK signal.
132
BPSK Receiver with Carrier Recovery
Circuits
133
The received signal is squared to generate a DC signal and an
amplitude varying sinusoid at twice the carrier frequency.
The DC signal is filtered out using a bandpass filter with center
frequency tuned to 2f
c
.
A frequency divider is used to recreate the waveform
cos(2f
c
t+).
The output of the multiplier is applied to an integrate and dump
circuit which forms the low pass filter segment of a BPSK
detector.
If the transmitter and receiver pulse shapes are matched, then the
detection will be optimum.
A bit synchronizer is used to facilitate sampling of the integrator
output precisely at the end of each bit period.
Operations of BPSK Receiver with Carrier
Recovery Circuits
134
Differential Phase Shift Keying (DPSK)
Differential PSK is a noncoherent form of phase shift keying
which avoids the need for a coherent reference signal at the
receiver.
1 k k k
d m d

=
135
Block Diagram of DPSK Receiver
136
Quadrature Phase Shift Keying (QPSK)
137
Spectrum of QPSK Signals
( )
( )
( )
( )
( )
2 2
sin sin
c s c s
QPSK b
c c
f f T f f T
P f E
f f T f f T

= +

138
Block Diagram of a QPSK Transmitter
139
Block Diagram of a QPSK Receiver
140
Offset QPSK (OQPSK)
For QPSK, the occasional phase shift of radians can cause the
signal envelope to pass through zero for just an instant.
The amplification of the zero-crossings brings back the filtered
sidelobes since the fidelity of the signal at small voltage levels is
lost in transmission.
To prevent the regeneration of sidelobes and spectral widening, it
is imperative that QPSK signals that use pulse shaping be
amplified only using linear amplifiers, which are less efficient.
A modified form of QPSK, called offset QPSK (OQPSK) or
staggered QPSK is less susceptible to these deleterious effects
and supports more efficient amplification.
OQPSK ensures there are fewer baseband signal transitions.
Spectrum of an OQPSK signal is identical to that of QPSK.
141
Offset QPSK (OQPSK)
The time offset waveforms that are applied to the in-phase and
quadrature arms of an OQPSK modulator. Notice that a half-
symbol offset is used.
142
/4-DQPSK
143
Generic /4-DQPSK Transmitter
144
/4-DQPSK Baseband Differential
Detector
145
Detection of Binary Signals in Gaussian
Noise
146
Digital Demodulation Techniques
Coherent detection: Exact replicas of the possible arriving
signals are available at the receiver. This means that the
receiver has exact knowledge of the carrier waves phase
reference, in which case we say the receiver is phase-locked to
the transmitter. Coherent detection is performed by cross-
correlating the received signal with each one of the replicas,
and then making a decision based on comparisons with pre-
selected thresholds.
Non-coherent detection: Knowledge of the carrier waves
phase is not required. The complexity of the receiver is
thereby reduced but at the expense of an inferior error
performance, compared to a coherent system.
147
Correlation Demodulator
148
Matched Filter Demodulator
149
Inter-Symbol Interference (ISI)
150
Inter Symbol Interference (ISI)
Inter-Symbol Interference (ISI) arises because of
imperfections in the overall frequency response of the
system. When a short pulse of duration T
b
seconds is
transmitted through a band-limited system, the
frequency components constituting the input pulse
are differentially attenuated and differentially delayed
by the system. Consequently, the pulse appearing at
the output of the system is dispersed over an interval
longer than T
b
seconds, thereby resulting in inter-
symbol interference.
Even in the absence of noise, imperfect filtering and
system bandwidth constraints lead to ISI.
151
Nyquist Channels for Zero ISI
The Nyquist channel is not physically realizable since it
dictates a rectangular bandwidth characteristic and an infinite
time delay.
Detection process would be very sensitive to small timing
errors.
Solution: Raised Cosine Filter.
152
Raised Cosine Filter
0
0
0
0
0
0
0
0 2
: Factor Off - Roll
: Bandwidth Excess
2
1
for
2 for
2 for

0
)
2
4
( cos
1
) (
W
W W
r
W W
T
W
W f
W f W W
W W f
W W
W W f
f H
=
>
< <
<
+
=

153
Raised Cosine Filter Characteristics
154
Raised Cosine Filter Characteristics
155
Equalization
In practical systems, the frequency response of the
channel is not known to allow for a receiver design that
will compensate for the ISI.
The filter for handling ISI at the receiver contains
various parameters that are adjusted with the channel
characteristics.
The process of correcting the channel-induced distortion
is called equalization.
156
Equalization
157
Introduction to RAKE Receiver
Multiple versions of the transmitted signal are seen at
the receiver through the propagation channels.
Very low correlation between successive chips is in
CDMA spreading codes.
If these multi-path components are delayed in time
by more than a chip duration, they appear like
uncorrelated noise at a CDMA receiver.
Equalization is
NOT necessary
Combine
Coherently
158
Introduction to RAKE Receiver
To utilize the advantages of diversity techniques,
channel parameters are necessary to be estimated.
Arrival time of each path, Amplitude, and Phase.
Maximal Ratio Combiner (MRC): The combiner that
achieves the best performance is one in which each
output is multiplied by the corresponding complex-
valued (conjugate) channel gain. The effect of this
multiplication is to compensate for the phase shift in the
channel and to weight the signal by a factor that is
proportional to the signal strength.
159
Maximum Ratio Combining (MRC)
MRC: G
i
=A
i
e
-jq
i
Coherent Combining
Channel Estimation
Best Performance
G
1
Receiver
G
2
G
L
160
1
2
2 2
,
1
2
2
1
2
2
2
,
1
2
2
,
1 1
,
Received Envelope:
Total Noise Power:
SNR:
2
2
Since
L
L l l
l
L
n l n l
l
L
l l
l
L
L
L
n
l n l
l
L L
l
l l l n l
l l
n l
r G r
G
G r
r
SNR
G
r
G r G
=
=
=
=
= =
=
=
= =

161
*
,
*
,
1 1
2
,
2
1
2
,
2
1 1
2
,
2
,
1 1
2
,
2
,
2
1
@ branches all from SNRs of Sum SNR Output
: hold equality With
2
1
2
1
: Inequality s Chebychev'
l l
l n
l
l n l
L
l
l
L
l
l n
l
L
l
l n l
L
l
L
l
l n
l
l n l
L
L
l
L
l
l n
l
l n l
L
l
l l
r G
r
k G
SNR
r
G
r
G
SNR
r
G r G
=
=
= =

= =
=
= =
= = =
162
Example of RAKE Receiver Structure
163
Advantages of RAKE Receiver
Consider a receiver with only one finger:
Once the output of a single correlator is corrupted by
fading, large bit error is expected.
Consider a RAKE receiver
If the output of a single correlator is corrupted by fading,
the others may NOT be.
Diversity is provided by combining the outputs
Overcome fading
Improve CDMA reception

3G 04 DigitalComm

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

3G 04 DigitalComm

Uploaded by

Copyright:

Available Formats

Wireless Information Transmission System Lab.

National Sun Yat-sen University Institute of Communications Engineering

You might also like