Digital Data Transmission

ECE 457
Spring 2005
 Analog vs. Digital
x(t)
 Analog signals
 Value varies continuously
t
 Digital signals x(t)
 Value limited to a finite set

t
 Binary signals x(t) 1 1 1
 Has at most 2 values
 Used to represent bit values 0 T 0 0 0
 Bit time T needed to send 1 bit
 Data rate R=1/ T bits per second t
 Information Representation
• Communication systems convert information into
a form suitable for transmission
• Analog systemsAnalog signals are modulated
(AM, FM radio)
• Digital system generate bits and transmit digital
signals (Computers)
• Analog signals can be converted to digital signals.
 Digital Data System

Figure 7-1 Block diagram of a digital data system. (a) Transmitter.

(b) Receiver.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 Components of Digital
Communication
• Sampling: If the message is analog, it’s converted to
discrete time by sampling.
(What should the sampling rate be ?)
• Quantization: Quantized in amplitude.
Discrete in time and amplitude
• Encoder:
– Convert message or signals in accordance with a set of rules
– Translate the discrete set of sample values to a signal.
• Decoder: Decodes received signals back into original
message
 Different Codes
0 1 1 0 1 0 0 1
 Performance Metrics
• In analog communications we want, mˆ (t )  m(t )
• Digital communication systems:
– Data rate (R bps) (Limited) Channel Capacity
– Probability of error Pe
– Without noise, we don’t make bit errors
– Bit Error Rate (BER): Number of bit errors that occur
for a given number of bits transmitted.
• What’s BER if Pe=10-6 and 107 bits are transmitted?
 Advantages
• Stability of components: Analog hardware
change due to component aging, heat, etc.
• Flexibility:
– Perform encryption
– Compression
– Error correction/detection
• Reliable reproduction
 Applications
• Digital Audio • Digital Audio
Transmission Recording
• Telephone channels • LP vs. CD
• Lowpass • Improve fidelity
filter,sample,quantize (How?)
• 32kbps-64kbps • More durable and
(depending on the don’t deteriorate with
encoder) time
 Baseband Data Transmission

Figure 7-2
System model and waveforms
for synchronous baseband
digital data transmission.
(a) Baseband digital data
communication system.
(b) Typical transmitted
sequence. (c) Received
sequence plus noise.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
• Each T-second pulse is a bit.
• Receiver has to decide whether it’s a 1 or 0
( A or –A)
• Integrate-and-dump detector
• Possible different signaling schemes?
 Receiver Structure

Figure 7-3 Receiver structure and integrator output. (a) Integrate-and-

dump receiver. (b) Output from the integrator.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 Receiver Preformance
• The output of the integrator:
t 0 T

V  [ s(t )  n(t )]dt

t0

 AT  N A is sent

 AT  N A is sent
t 0 T

• N   n(t )dt is a random variable.

t0

• N is Gaussian. Why?
 Analysis
t 0 T t 0 T

E[ N ]  E[ 
t0
n(t )dt ]   E[n(t )]dt  0
t0

Var [ N ]  E[ N 2 ]  E 2 [ N ]
 E[ N 2 ] Why ?
 t 0 T 
2

 
 E   n(t )dt  
 t0
  
t 0 T t 0 T

   E[n(t )n( s)]dtds

t0 t0
t 0 T t 0 T
N0
  
t0 t0
2
 (t  s )dtds Why ?(White noise is uncorrelated!)

N 0T
• Key 
Point2
– White noise is uncorrelated
 Error Analysis
• Therefore, the pdf of N is:
 n 2 /( N 0T )
e
f N ( n) 
N 0T

• In how many different ways, can an error

occur?
 Error Analysis
• Two ways in which errors occur:
– A is transmitted, AT+N<0 (0 received,1 sent)
– -A is transmitted, -AT+N>0 (1 received,0 sent)

Figure 7-4 Illustration of error probabilities for binary signaling.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
  AT
e  n 2 / N 0T 2 A2T 
• P ( Error | A)   dn  Q 
N 0T  N0 
  

• Similarly,

e  n 2 / N 0T  2 A2T 
P ( Error |  A)   dn  Q 
N 0T  N0 
AT  

• The average probability of error:

PE  P ( E | A) P ( A)  P ( E |  A) P (  A)
 2 A2T 
 Q 
 N0 
 
• Energy per bit:
t 0 T

Eb   
2 2
A dt A T
t0

• Therefore, the error can be written in terms

of the energy.
• Define
A2T Eb
z 
N0 N0
• Recall: Rectangular pulse of duration T
seconds has magnitude spectrum
AT sinc (Tf )
• Effective Bandwidth: Bp  1/ T

• Therefore,
A2
z
N0 Bp

• What’s the physical meaning of this

quantity?
Probability of Error vs. SNR

Figure 7-5
PE for antipodal baseband
digital signaling.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 Error Approximation
• Use the approximation
u 2 / 2
e
Q(u )  , u  1
u 2
 2 A2T  e z
PE  Q  , z  1
 N0  2 z
 
 Example
• Digital data is transmitted through a
baseband system with N 0  107W / Hz , the
received pulse amplitude A=20mV.
a)If 1 kbps is the transmission rate, what is
probability of error?
1 1
Bp   3  103
T 10
A2 400  106
SNR  z   7  400  10 2  4
N 0 B p 10  10 3

ez
PE   2.58 10 3
2 z
b) If 10 kbps are transmitted, what must be
the value of A to attain the same
probability of error?
A2 A2 3
z  7  4  A 2
 4  10  A  63.2mV
N 0 B p 10 10 4

• Conclusion:
Transmission power vs. Bit rate
 Binary Signaling Techniques

Figure 7-13
Waveforms for ASK, PSK, and
FSK modulation.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 A SK , P SK , an d F SK
 A m p litu d e S h ift K e y in g ( A S K ) 1 0 1 1
m (t)
 A cos( 2  f c t ) m ( nT b )  1
s ( t )  m ( t ) A c cos( 2  fct)   c
 0 m ( nT b )  0
A M M o d u la tio n
 P h a s e S h ift K e y in g ( P S K )
1 0 1 1
m (t)
 A c cos( 2  f c t ) m ( nT b )  1
s ( t )  A c m ( t ) cos( 2  fct)  
 A c cos( 2  f c t   ) m ( nT b )   1
P M M o d u la tio n
 F re q u e n c y S h ift K e y in g 1 0 1 1
 A cos( 2  f1t ) m ( nT b )  1
s (t)   c
 A c cos( 2 f2t) m ( nT b )   1
F M M o d u la tio n
 Amplitude Shift Keying (ASK)
• 00
• 1Acos(wct)
• What is the structure of the optimum
receiver?
Receiver for binary signals in
noise

Figure 7-6 A possible receiver structure for detecting binary signals in

white Gaussian noise.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 Error Analysis
• 0s1(t), 1s2(t) in general.
• The received signal:
y (t )  s1 (t )  n(t ), t0  t  t0  T
OR
y (t )  s2 (t )  n(t ), t0  t  t0  T
• Noise is white and Gaussian.
• Find PE
• In how many different ways can an error occur?
 Error Analysis (general case)
• Two ways for error:
» Receive 1 Send 0
» Receive 0Send 1

• Decision:
» The received signal is filtered. (How does this
compare to baseband transmission?)
» Filter output is sampled every T seconds
» Threshold k
» Error occurs when:
v(T )  s01 (T )  n0 (T )  k
OR
v(T )  s02 (T )  n0 (T )  k
• s01 , s02 , n0 are filtered signal and noise terms.
• Noise term: n 0(t ) is the filtered white Gaussian
noise.
• Therefore, it’s Gaussian (why?)
• Has PSD: N0 2
S n0 ( f )  H( f )
2
• Mean zero, variance?
• Recall: Variance is equal to average power of the
noise process

N0
 2
2
  2
H ( f ) df
• The pdf of noise term is:
 n 2 / 2 2 0
e
f N ( n) 
2 2
• Note that we still don’t know what the filter is.
• Will any filter work? Or is there an optimal one?
• Recall that in baseband case (no modulation), we
had the integrator which is equivalent to filtering
with 1
H( f ) 
j 2f
• The input to the thresholder is:
V  v(T )  s01 (T )  N
OR
V  v(T )  s02 (T )  N

• These are also Gaussian random variables; why?

• Mean: s01 (T ) OR s02 (T )
• Variance: Same as the variance of N
 Distribution of V
• The distribution of V, the input to the
threshold device is:

Figure 7-7 Conditional probability density functions of the filter output

at time t = T.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 Probability of Error
• Two types of errors:
 [ v  s01 (T )]2 / 2 2
e  k  s01 (T ) 
P ( E | s1 (t ))   dv  Q 
k 2 2   
k [ v  s02 (T )]2 / 2 2
e  k  s02 (T ) 
P ( E | s2 (t ))  
 2 2
dv  1  Q
 



• The average probability of error:

1 1
PE  P[ E | s1 (t )]  P[ E | s2 (t )]
2 2
• Goal: Minimize the average probability of
errror
• Choose the optimal threshold
• What should the optimal threshold, kopt be?
• Kopt=0.5[s01(T)+s02(T)]
•  s02 (T )  s01 (T ) 
P  Q 
E
 2 
 Observations
• PE is a function of the difference between the two
signals.
• Recall: Q-function decreases with increasing
argument. (Why?)
• Therefore, PE will decrease with increasing
distance between the two output signals
• Should choose the filter h(t) such that PE is a
minimummaximize the difference between the
two signals at the output of the filter
 Matched Filter
• Goal: Given s1 (t ), s2 (t ) , choose H(f) such
s (T )  s (T )
that d 02 01


is maximized.
• The solution to this problem is known as
the matched filter and is given by:
h0 (t )  s2 (T  t )  s1 (T  t )

• Therefore, the optimum filter depends on

the input signals.
 Matched filter receiver

Figure 7-9 Matched filter receiver for binary signaling in white

Gaussian noise.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 Error Probability for Matched
Filter Receiver
• Recall d 
PE  Q 
 2of
• The maximum value  the distance,

2
( E1  E2  2 E1 E2 12 )
2
d max 
• E1 is the energy of theNfirst
0
signal.
• E2 is the energy of the second signal.

t 0 T


2
E1  s1 (t ) dt
t0
t 0 T


2
E2  s2 (t ) dt
t0

1
12 
E1 E2 
 s (t ) s
1 2 (t ) dt
• Therefore,

 E  E  2 E E  1/ 2
 
PE  Q  1 2 1 2 12  
 2N0 
 
 

• Probability of error depends on the signal energies

(just as in baseband case), noise power, and the
similarity between the signals.
• If we make the transmitted signals as dissimilar as
possible, then the probability of error will decrease (
)
12  1
 ASK
s1 (t )  0, s2 (t )  A cos(2f c t )

• The matched filter: A cos(2f ct )

• Optimum Threshold: 4 A T 1 2

• Similarity between signals?

• Therefore, P  Q 4ANT   Q z 
2

E
 0 

• 3dB worse than baseband.

 PSK
1 1
s1 (t )  A sin(2f ct  cos m), s2 (t )  A sin(2f ct  cos m)
• Modulation index: m (determines the phase
jump)
• Matched Filter:  2 A 1  m cos(2f t )
2
c

• Threshold: 0
• Therefore, P  Q( 2(1  m ) z )
E
2

• For m=0, 3dB better than ASK.

 Matched Filter for PSK

Figure 7-14 Correlator realization of optimum receiver for PSK.

Principles of Communications, 5/E by Rodger Ziemer and William Tranter

Copyright © 2002 John Wiley & Sons. Inc. All rights reserved.
 FSK
• s1 (t )  A cos(2f c t ), s2 (t )  A cos(2 ( f c  f )t )
• f  m
T
• Probability of Error: Q( z )
• Same as ASK
 Applications
• Modems: FSK
• RF based security and access control
systems
• Cellular phones