EGR 544 Communication Theory

7. Representation of Digitally Modulated Signals II

Z. Aliyazicioglu

Electrical and Computer Engineering Department

Cal Poly Pomona

Representation of Digital Modulation with Memory

Linear Digital Modulation with Memory

NRZ modulation is equivalent to

NRZ binary PAM or PSK. It is
memoryless system

NRZ1 NRZ1 is called differential encoding.

If A is pulse amplitude,
A change when data=1 and
Data 1 0 1 1 0 0 0 1 A doesnt change when data=0

bk = ak bk 1

where {ak} s binary information sequence, {bk} is the output

sequence, addition modulo 2

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 2

1
 Linear Digital Modulation with Memory
State diagram for the NRZ1 represents the encoder and modulator
operations. s(t) represents the waveform to transmit the binary
information Data s(t) for Data 1 and s(t) for Data 0
1/s(t) Input bit/channel symbol
0/-s(t) 0/s(t)
State diagram, It is also
called Markow Chain
S0=0 S1=1
1/-s(t)
State transition 1 0 0 1 State transition
matrix for input 0 T0 = T1 = matrix for input 1
0 1 1 0
0/-s(t) 0/-s(t) 0/-s(t) 0/-s(t)
S0=0

1/s(t) 1/s(t) 1/s(t) 1/s(t)

The trellis diagram for
1/-s(t) 1/-s(t) 1/-s(t) 1/-s(t) NRZ1
S1=1
0/s(t) 0/s(t) 0/s(t) 0/s(t)
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 3

Linear Digital Modulation with Memory

Delay Modulation Miller Code s1(t) s2(t)

t T t
NRZ 0
0 T
State S1 S2 S4 S3 S2 S4 S1 S4 S3 s3(t)
s4(t)
Miller 0 T
Code t
t
0 T
Data 1 0 1 1 0 0 0 1
s4(t)= -s1(t) for 0<t<T
S2
1/s2(t) 1/s2(t) s3(t)= -s2(t) for 0<t<T
1/s3(t)
0/s1(t) Basic Waveform for Miller Code
S1 S3
0/s1(t) 0/s4(t)

0/s4(t) 1/s3(t) State Diagram for Miller Code

S4
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 4

2
 Linear Digital Modulation with Memory
State transition matrix S2
1/s2(t) 1/s2(t)
T T 1/s3(t)
S1 (t ) S1 (t 1) 0 0 0 1 0/s1(t)
S (t ) S (t 1) 0 0 0 1
S1 S3
2 = 2 0/s1(t) 0/s4(t)
S3 ( t ) S3 (t 1) 1 0 0 0
0/s4(t) 1/s3(t)
S4 (t ) S4 (t 1) 1 0 0 0 S4

p( S1 | S1 , I = 0) p( S2 | S1 , I = 0) p( S3 | S1 , I = 0) p( S4 | S1 , I = 0)
p( S | S , I = 0) p( S2 | S2 , I = 0) p( S3 | S2 , I = 0) p( S4 | S2 , I = 0)
T0 = 1 2
p ( S1 | S3 , I = 0) p ( S2 | S3 , I = 0) p ( S3 | S3 , I = 0) p( S4 | S3 , I = 0)

p( S1 | S4 , I = 0) p( S2 | S4 , I = 0) p( S3 | S4 , I = 0) p( S4 | S4 , I = 0)
0 0 0 1 0 1 0 0
0 0 1 0
0 0 1 0
T0 = T1 =
1 0 0 0 Similarly 0 1 0 0

1 0 0 0 0 0 1 0
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 5

Linear Digital Modulation with Memory

State or channel Symbol) transition matrix

Let pij=P(Si|Sj). Then the matrix P=[pij] is called the transition

probability matrix which is give for delay modulation as

p11 p12 p13 p14

p p22 p23 p24
P = 21
P = P( I = 0)T0 + P( I = 1)T1
p31 p32 p33 p34

p41 p42 p43 p44

Example: The transition matrix with equal likely symbols P(0)=P(1)=1/2

0 0 0 1 0 1 0 0 0 1/ 2 0 1/ 2
0 0 1 1 0 0 1 0 0 0 1/ 2 1/ 2
1 0
P= + =
2 1 0 0 0 2 0 1 0 0 1/ 2 1/ 2 0 0

1 0 0 0 0 0 1 0 1/ 2 0 1/ 2 0
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 6

3
 Non-Linear Modulation with Memory
Continuous Phase FSK (CPFSK)
First, lets look at the FSK

s-(M-1)(t) Modulation
s-(M-2)(t) s0(t), s1(t),..

selector
Channel
s(M-1)(t)

Channel
Encoder
01011011101 01111001101
Source Output

FSK
2 1
sm (t ) = cos 2 f c t + 2 ( f I n )t
T 2 where I n = 1, 3,... ( M 1)

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 7

Non-Linear Modulation with Memory

FSK signal is generated by shifting carrier frequency to represent the
digital information.

1
fn = f I n , I n = 1, 3,..., ( M 1)
2
Needs to have M=2k separate oscillator to to represent each
signal. The sudden switching from one signal frequency to
another needs large bandwidth for transmission

To avoid the problem, continuous-phase FSK is used

Singe carrier whose frequency is changed continuously

This types of FSK signal has memory because the phase of the
carriers is constrained to be continuous

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 8

4
 Continuous Phase FSK (CPFSK)

To represent CPFSK, lets use PAM signal for each k-bits block

d (t ) = I n g (t nT ) is called delta function

n

Where In represent the amplitude values 1, 3,, (M-1) and

each of them maps k-bit blocks of information sequence
g(t) is rectangular pulse, amplitude 1/2T and duration T second.

d(t) signal is used to frequency modulate the carrier and the

carrier -modulated signal is
2
cos 2 f c t + 4 Tf d d ( ) d
t
s (t ) = fd: peak frequency deviation
T

t
The function

d ( ) d Not consists of jump, this makes the result
phase shift with memory

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 9

Continuous Phase FSK (CPFSK)

For simplicity

Let g(t) be a rectangular pulse of amplitude 1/2T at [0.T]

2
s (t ) = cos [2 f ct + (t; I)]
T

where
t
(t; I) = 4 Tf d d ( ) d

t
= 4 Tf d I n g ( nT ) d

n
n 1 T (t nT )
= 4 Tf d I k + In , t [nT ,(n + 1)T ]
k = 2T 2T
n 1
= 2 f d T I
k =
k + 2 f d (t nT ) I n

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 10

5
 Continuous Phase FSK (CPFSK)

The phase of carrier in nT t (n+1)T is

n 1
(t; I) = 2 Tf d I
k =
k + 2 f d (t nT ) I n Special case of CPM
explained in next slide
= n + 2 hI n q(t nT )

where h = 2Tf d Modulation index

n 1
n = h I k Represents the accumulator (memory) up
k = to time (n-1)T

(t < 0) 1/2
0
t/2T
q(t ) = t / 2T (0 t T )
1/ 2 (t > 0)
0 T

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 11

Continues-phase Modulation (CPM)

The carrier phase of continuous-phase modulated signal is
n
(t; I) = 2 I h q(t kT )
k =
k k nT t ( n + 1)T

where
{Ik} 1, 3,, (M-1) sequence of M-ary information symbols
{hk} is a sequence of modulation index
q(t) is normalized waveform shape.

When hk is not fixed, the CPM signal is called multi-h

The normalized waveform q(t) can be represented as

t
q(t ) = g ( ) d
0

CPM signal is called full response CPM, if g(t)=0 for t>T

CPM signal is called partial response CPM, if g(t)0 for some t>T

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 12

6
 Continues-phase Modulation (CPM)
Example of full response CPM; Some shapes of g(t) and q(t)
g(t) q(t)
1/2T 1/2

t t
T T

g(t) q(t)
1/2
1/4T

t
2T t
2T

1
0 1 LT
g (t ) = 2 LT
0 (o.w)
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 13

Continues-phase Modulation (CPM)

Example of partial response CPM;Some shapes of g(t) and q(t)
1 2 t q(t)
g (t ) = (1 cos )
2T T
1/T 1/2
t 1 2 t
q( t ) = sin
2T 4 T
t t
T T
1 t
g (t ) = (1 cos )
4T T q(t)
1/2
1/2T t 1 t
q( t ) = sin
4T 4 T
t
2T t
2T
1 2 t
(1 cos ) 0 1 LT
g (t ) = 2 LT LT
0 (o.w)
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 14

7
 Continues-phase Modulation (CPM)

Representation of Continuous-Phase Modulation

Phase trajectory of phase tree trellis
Phase trellis
Phase cylinder
Phase state trellis
Phase state diagram

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 15

Continues-phase Modulation (CPM)

Example 1.
The case of Binary
CPFSK with In= 1
and g(t) is a full
response rectangular
function.
The set of phase
trajectories starting
t=0

n 1
(t; I) = h I k + 2 hI n q(t nT )
k =

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 16

8
 Continues-phase Modulation (CPM)

Example 2:
The case of Quaternary
CPFSK with In= 1
3 and full response
rectangular function,
the set of phase
trajectories starting
t=0

n 1
(t; I) = h I k + 2 hI n q(t nT )
k =

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 17

Continues-phase Modulation (CPM)

Example:
A phase trajectories generated by the sequence In=(1,-1,-1,-
1,1,1,01,1) for the partial response
1 2 t
g (t ) = (1 cos ) Raised cosine of length 3T
6T 3T

Phase trajectories for binary CPFSK (dashed) and binary, partial response
CPM based on raised cosine pulse of length 3T (solid). [From Sundberg
(1986), 1986 IEEE.]

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 18

9
 Continues-Phase Modulation (CPM)
Phase State trellis
Simple way to represent the phase trajectories is concern only
those phase values at t=nT. Range from =0 to =.
(nT , I) s = {0, h , 2h ,3h ,...}

Example: For a full response CPM and h=m/p

m 2 m ( p 1) m For m is even
s = 0, , ,...,
p p p There are p terminal phase state
m 2 m (2 p 1) m For m is odd
s = 0, , ,...,
p p p There are 2p terminal phase state
The maximum number of phase state is

pM L1 m even M is alphabet size

St = L 1 L is an integer # extends the pulse
2 pM m odd
shape
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 19

Continues-phase Modulation (CPM)

Example: The phase state of the binary CPFSK (full response)
with h=1/2 and St=4

Difference between phase state trellis and phase trellis is:

The connection between state are made by drawing straight lines for phase state
trellis. This is not true for phase trajectories from one state to another
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 20

10
 Continues-phase Modulation (CPM)

Phase state diagram

Only possible phase states and their transitions are displayed
Only concern those phase values at t=nT
Time not appear explicitly as a variable

0 = 0
1 =
2

3
3 = 2 =
2

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 21

Minimum-shift Keying (MSK)

MSK is special case of CPFSK and modulation index h=1/2.
The phase of the carrier in the interval nT t (n+1)T is

n 1 g(t)
1
(t; I) =
2

k =
I k + I n q(t nT ) 1/2T

1 t nT
= n + In ( ), nT t ( n + 1)T t
2 T nT (n+1)T
The modulated carrier signal is

1 t nT
s(t ) = A cos 2 f ct + n + I n
2 T
1 1
= A cos 2 f c + I n t n I n + n
4T 2

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 22

11
 Minimum-shift Keying (MSK)
The binary CPFSK signal will have two frequencies in the
interval nT t (n+1)T

1 1
f1 = f c f2 = fc +
4T 4T

The binary CPFSK signal can be also written as

1
si (t ) = A cos 2 f i t + n + n ( 1)i 1 , i = 1, 2
2
The frequency separation f = f 2 f1 = 1/ 2T

This the minimum frequency separation that is necessary to ensure

the orthogonality of the signal s1(t) and s2(t) over a signaling
interval of the length T. That is why binary CPFSK with h=1/2 is
MSK

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 23

Minimum-shift Keying (MSK)

Also, MSK map be represents as a form of four-phase PSK.

The equivalent low-pass digitally modulated signal is

v (t ) = A [ I 2 n g (t 2nT ) + jI 2 n +1 g (t (2n + 1)T )]
n =

t
where sin , t [0, 2T )
g (t ) = 2T
0, o.w.

Viewed as a four-phase PSK signal with pulse shape is one-half

cycle of sinusoidal. The even numbered binary valued symbols
I2n of the information sequence are transmitted via cos of the
carrier. The odd-numbered symbols {I2n+1} are transmitted via
the sin carrier

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 24

12
 Minimum-shift Keying (MSK)
Transmission rate for each is 1/2T bits/ss, combine transmission
rate will be 1/T bit/s.
MSK

s(t ) = A [ I 2 n g (t 2nT ) cos 2 f ct + I 2 n +1 g (t (2n + 1)T )sin 2 f ct ]
n =

s(t), a constant amplitude and frequency modulated signal

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 25

Minimum-shift Keying (MSK)

A [ I 2 n g (t 2nT ) cos 2 f ct ]
n =

In phase signal component

A [ I 2 n +1 g (t (2n + 1)T )sin 2 f ct ]
n =

Quadrature signal component

s(t), a constant amplitude and

frequency modulated signal

Frequency at [nT ,( n + 1)T )

1
fc + In
4T
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 26

13
 Comparison of MSK and QPSK

MSK
s(t ) = A [ I 2 n g (t 2nT ) cos 2 f ct + I 2 n +1 g (t (2n + 1)T )sin 2 f ct ]
n =

t
sin , t [0, 2T ) Continuous phase
where g (t ) = 2T
0, o.w.

Offset QPSK
s(t ) = A [ I 2 n g (t 2nT ) cos 2 f c t + I 2 n +1 g (t (2n + 1)T )sin 2 f c t ]
n =

where 1, t [9, 2T ) Possibly 90 degree phase jump

g (t ) =
0, o.w. at each T
QPSK

I 2 n + I 2 n +1 I I
s (t ) =
n = 2
g (t 2nT ) cos 2 f ct + 2 n 2 n +1 g (t 2nT )sin 2 f ct
2
where Possibly 90 or 180 degree phase
Cal Poly Pomona Electrical & Computer jump at each
Engineering Dept.2T EGR 544-7 27

Minimum-shift Keying (MSK)

QPSK
-90 phase 90 phase
shift shift
1 2
( I , I , I , I ) = ( +1, 1, +1, 1)
MSK : 0 1 2 3
f c = 1.25( f1 = 1; f 2 = 1.5)

( I , I , I , I ) = ( +1, +1, 1, 1)
OQPSK : 0 1 2 3
f c = 1.5
90 phase 90 phase
shift shift
(I , I , I , I , I , I ) = (+1, +1, 1, 11, +1)
QPSK : 0 1 2 3 4 5
fc =1.5

180 phase -90 phase

shift shift

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 28

14
 1 t T
11 (t; I) = 1 + I1 ( ) for T t nT
2 T
1 1 t T
= 0 + I 0 + I1 ( ) (2T ; I) (T ; I) =
2 2 T 2
1 1
= 0 + (t 1)
2 2
1
= t + + 0
2
1 t 2T
(t; I) = 2 + I 2 ( ) for 2T t 3T
2 2 T (3T ; I) (2T ; I) =
2
1 1 1 t 2T
= 0 + I 0 + I1 + I 2 ( )
2 2 2 T

= 0 + + (t 2)
2 2 2

= t + 0
2
Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 29

Signal space diagram of CPM

Continues phase signal can not be represented by discrete points in
signal space, like PAM, PSK
It can be described by the trajectories from one phase state to another

Here is signal phase trajectories diagram for CPFSK signal for

h=1/4,h=1/3,h=1/2, and h=2/3

H=1/4 H=1/3 H=1/2 H=2/3

Cal Poly Pomona Electrical & Computer Engineering Dept. EGR 544-7 30

15