Digital Communication

 Power Efficiency →  p : ability of a modulation

technique to preserve the quality of digital messages at
low power levels (low SNR)
 Specified as Eb / No @ some BER (e.g. 10-5) where Eb : energy/bit
and No : noise power/bit
 Tradeoff between fidelity and signal power →
BER ↑ as Eb / No ↓

1
 Bandwidth Efficiency →  B : ability of a modulation
technique to accommodate data in a limited BW
R
 B  bps/Hz R : data rate B: RF BW
B
 Tradeoff between data rate and occupied BW
→ as R ↑, then BW ↑
 For a digital signal :

2
 each pulse or “symbol” having m finite states
represents n = log2 m bits/symbol →
 e.g. m = 0 or 1 (2 states) → 1 bit/symbol (binary)
 e.g. m = 0, 1, 2, 3, 4, 5, 6, or 7 (8 states) → 3 bits/symbol

3
 Implementation example: A system is changed from binary
to 2-ary.
 Before: "0" = - 1 Volt, "1" = 1 Volt
 Now
"0" = - 1 Volt, "1" = - 0.33 volts, "2" = 0.33 Volts, "3" = 1 Volt

 What would be the new data rate compared to the old data
rate if the symbol period where kept constant?

 In general, called M-ary keying

4
 Maximum BW efficiency → Shannon’s Theorem
 Most famous result in communication theory.

 where

 B : RF BW
 C : channel capacity (bps) of real data (not retransmissions
or errors)
 To produce error-free transmission, some of the bit rate will
be taken up using retransmissions or extra bits for error
control purposes.
 As noise power N increases, the bit rate would still be the
same, but max  Bmax decreases.

5
 So

 note that C ∝ B (expected) but also C ∝ S / N

 an increase in signal power translates to an increase in channel
capacity
 lower bit error rates from higher power → more real data
 large S / N → easier to differentiate between multiple signal states
(m) in one symbol ∴ n ↑
 max  B is fundamental limit that cannot be achieved in
max

practice

6
 People try to find schemes that correct for errors.

 People are starting to refer to certain types of codes as

“capacity approaching codes”, since they say they are
getting close to obtaining Cmax.

7
8
9
 Fundamental tradeoff between  B and  p (in general)
 If  B improves then  p deteriorates (or vice versa)
 May need to waste more power to get a better data rate.
 May need to use less power (to save on battery life) at the
expense of a lower data rate.
  p vs.  B is not the only consideration.
 Use other factors to evaluate → complexity, resistance to MRC
impairments, etc.

10
 Bandwidth Specifications
 Many definitions depending on application → all use Power
Spectral Density (PSD) of modulated bandpass signal

 W (f)2 
SW ( f )  lim  T 
T   T 
 

 Many signals (like square pulses) have some power at all

frequencies.

11
  B’ : half-power (-3 dB) BW
 B” : null-to-null BW
 B’” : absolute BW →
range where PSD > 0
 FCC definition of occupied BW → BW contains 99% of
signal power

12
Amplitude Shift Keying (ASK)

Baseband
Data

1 0 0 1 0
ASK
modulated
signal

Acos(t) Acos(t)

 Pulse shaping can be employed to remove spectral spreading

 ASK demonstrates poor performance, as it is heavily affected by noise,
fading, and interference

13
Frequency Shift Keying (FSK)

Baseband
Data

1 0 0 1
BFSK
modulated
signal

f1 f0 f0 f1

where f0 =Acos(c-)t and f1 =Acos(c+)t

Example: The ITU-T V.21 modem standard uses FSK

FSK can be expanded to a M-ary scheme, employing multiple frequencies
as different states

14
Phase Shift Keying (PSK)

Baseband
Data

1 0 0 1
BPSK
modulated
signal

s1 s0 s0 s1

where s0 =-Acos(ct) and s1 =Acos(ct)

Major drawback – rapid amplitude change between symbols due to phase

discontinuity, which requires infinite bandwidth. Binary Phase Shift Keying
(BPSK) demonstrates better performance than ASK and BFSK
BPSK can be expanded to a M-ary scheme, employing multiple phases and
amplitudes as different states

15
III. Geometric Representation of Modulation
Signal
 Geometric Representation of Modulation Signals -
Constellation Diagrams
 Graphical representation of complex ( A & θ) digital
modulation types
 Provide insight into modulation performance
 Modulation set, S, with M possible signals

 Binary modulation → M = 2 → each signal = 1 bit of information

 M-ary modulation → M > 2 → each signal > 1 bit of information

16
 Example: Binary Phase Shift Keying (BPSK)

17
 Phase change between bits → Phase shifts of 180° for each
bit.
 Note that this can also be viewed as AM with +/- amplitude
changes

 Dimension of the vector space is the # of basis signals

required to represent S.

18
 Plot amplitude & phase of S in vector space :

19
 Constellation diagram properties :
1) Distance between signals is related to differences in
modulation waveforms
 Large distance → “sparse” → easy to discriminate → good BER @
low SNR (Eb / No )
 From above, noise of -2 added to would make the received
signal look like s2(t) → error.
 From , noise of > - would make the result closer to -
and would make the decoder choose s2(t) → error.
∴ Above example is Power Efficient (related to density with respect
to # states/N)

20
2) Occupied BW ↓ as # signal states ↑
 If we can represent more bits per symbol, then we need
less BW for a given data rate.
 Small separation → “dense” → more signal states/symbol
→ more information/Hz !!
∴ Bandwidth Efficient

21
Linear Modulation Techniques:
Digital modulation can be broadly classified as:
1. Linear (change Amplitude or phase)
2. Non linear modulation techniques (change frequency).
Linear Modulation Techniques:
• The amplitude /phase of the transmitted signal s(t), varies linearly with the modulating
digital signal, m(t).
• These are bandwidth efficient (because it doesn’t change frequency) and hence are
very attractive for use in wireless communication systems where there is an
increasing demand to accommodate more and more users within a limited spectrum.

22 of 82
Pros & Cons

• Linear Modulation schemes have very

good spectral efficiency,
•However, they must be transmitted
using linear RF amplifiers which have
poor power efficiency.

23 of 30
Note
 “Phase modulation” can be regarded as “amplitude”
modulation because it can really change “envelope”;

 Thus both of them belong to “linear modulation”!

24 of 30
IV. Linear Modulation Methods
 The amplitude /phase of the transmitted signal s(t), varies
linearly with the modulating digital signal, m(t).
 Performance is evaluated with respect to Eb / No

25
BPSK
 BPSK → Binary Phase Shift Keying

26
 Phase transitions force carrier amplitude to change
from “+” to “−”.
 Amplitude varies in time

27
BPSK RF signal BW

 Null-to-null RF BW = 2 Rb = 2 / Tb
 90% BW = 1.6 Rb for rectangular pulses
28
 Probability of Bit Error is proportional to the distance
between the closest points in the constellation.
 A simple upper bound can be found using the assumption
that noise is additive, white, and Gaussian.

 d is distance between nearest constellation points.

29
  Q(x) is the Q-function, the area under a normalized Gaussian
function (also called a Normal curve or a bell curve)

1  y2 / 2
Q( z )   e dy
z 2

 Here

30
 Demodulation in Rx
 Requires reference of Tx signal in order to properly
determine phase
 carrier must be transmitted along with signal
 Called Synchronous or “Coherent” detection
 complex & costly Rx circuitry
 good BER performance for low SNR → power efficient

31
32
DPSK
 DPSK → Differential Phase Shift Keying
 Non-coherent Rx can be used
 easy & cheap to build
 no need for coherent reference signal from Tx
 Bit information determined by transition between two
phase states
 incoming bit = 1 → signal phase stays the same as previous bit
 incoming bit = 0 → phase switches state

33
  If {mk} is the message, the output {dk} is as shown below.

 can also be described in modulo-2 arithmetic dk  mk  dk 1

 Same BW properties as BPSK, uses same amount of spectrum
 Non-coherent detection → all that is needed is to compare
phases between successive bits, not in reference to a Tx phase.
 power efficiency is 3 dB worse than coherent BPSK (higher
power in Eb / No is required for the same BER)

34
35
QPSK
 QPSK → Quadrature Phase Shift Keying

 Four different phase states in one symbol period

 Two bits of information in each symbol
Phase: 0 π/2 π 3π/2 → possible phase values
Symbol: 00 01 11 10

36
 Note that we choose binary representations so an
error between two adjacent points in the constellation
only results in a single bit error

 For example, decoding a phase to be π instead of π/2 will

result in a "11" when it should have been "01", only one bit
in error.

37
 Constant amplitude with four different phases
 remembering the trig. identity

38
39
 Now we have two basis functions
 Es = 2 Eb since 2 bits are transmitted per symbol
 I = in-phase component from sI(t).
 Q = quadrature component that is sQ(t).

40
QPSK RF Signal BW

 null-to-null RF BW = Rb = 2RS (2 bits / one symbol time) = 2 / Ts

 double the BW efficiency of BPSK → or twice the data rate in same
signal BW
41
 BER is once again related to the distance between
constellation points.

 d is distance between nearest constellation points.

42
43
 How does BER performance compare to BPSK?

 Why? same # of states per number of basis functions for

both BPSK and QPSK (2 states per one function or 4 states
per 2 functions)
 same power efficiency
(same BER at specified Eb / No)
 twice the bandwidth efficiency
(sending 2 bits instead of 1)

44
 QPSK Transmission and Detection Techniques

45
46