The Basics
[w/Hz]
Power spectral
Density
(flat => white)
Autocorrelation
Function
(uncorrelated)
likelihoods
pz (z | m1 )
s2
Eb
s1
0
1 (t )
Eb
s1 s 2 / 2
Pe (m1 ) Pe (m2 ) Q
N 0 / 2
PB PE (2) Q
2 Eb
N0
Bigger Picture
General structure of a communication systems
Noise
Source
SOURCE
Info.
Received
Transmitted
Received
info.
signal
signal
Transmitter
Receiver
Channel
User
Transmitter
Formatter
Source
encoder
Channel
encoder
Modulator
Receiver
Formatter
Source
decoder
Channel
decoder
Demodulator
Original
pulse
Regenerator receiver
Regenerated
pulse
Propagation distance
Voice
Data
A bit is a bit!
Media
Output
Linear system
Deterministic signals:
Random signals:
Ideal distortion less transmission:
All the frequency components of the signal not only
arrive with an identical time delay, but also are
amplified or attenuated equally.
Low-pass
Band-pass
Realizable filters:
RC filters
High-pass
Bandwidth of signal
Baseband versus bandpass:
Baseband
signal
Bandpass
signal
Local oscillator
Bandwidth dilemma:
Bandlimited signals are not realizable!
Realizable signals have infinite bandwidth!
Bandwidth of signal:
Approximations
Different
definition of bandwidth:
a) Half-power bandwidth
b) Noise equivalent bandwidth
c) Null-to-null bandwidth
(a)
(b)
(c)
(d)
(e)50dB
Textual
source info.
Analog
info.
Analog
info.
sink
Sample
Quantize
Format
Low-pass
filter
Textual
info.
Digital info.
Decode
Pulse
modulate
Encode
Bit stream
Pulse
waveforms
Demodulate/
Detect
Transmit
Channel
Receive
xs (t ) x (t ) x(t )
x (t )
x (t )
xs (t )
Frequency domain
X s ( f ) X ( f ) X ( f )
| X(f )|
| X ( f ) |
| Xs( f )|
LP filter
Nyquist rate
aliasing
Undersampling &Aliasing in
Time Domain
Note: correct reconstruction does not draw straight lines between samples.
Key: use sinc() pulses for reconstruction/interpolation
Sampling theorem
Analog
signal
Sampling
process
Pulse amplitude
modulated (PAM) signal
In practice need to sample faster than this because the receiving filter
will not be sharp.
Quantization
Amplitude quantizing: Mapping samples of a continuous
amplitude waveform to a finite set of amplitudes.
Out
In
Quantized
values
Encoding (PCM)
A uniform linear quantizer is called Pulse Code
Modulation (PCM).
Pulse code modulation (PCM): Encoding the
quantized signals into a digital word (PCM word or
codeword).
Each quantized sample is digitally encoded into
an l bits codeword where L in the number of
quantization levels and
Quantization error
Quantizing error: The difference between the input
and output of a quantizer
e(t ) x (t ) x(t )
y q (x)
AGC
x(t )
x (t )
x(t )
x (t )
x
e(t )
e(t )
x (t ) x(t )
Non-uniform quantization
It is done by uniformly quantizing the compressed signal.
At the receiver, an inverse compression characteristic, called
expansion is employed to avoid signal distortion.
compression+expansion
y C (x)
x(t )
companding
x
y (t )
y (t )
x (t )
y
x
Compress
Qauntize
Transmitter
Expand
Channel
Receiver
Baseband transmission
To transmit information thru physical channels, PCM
sequences (codewords) are transformed to pulses
(waveforms).
Each waveform carries a symbol from a set of size M.
Each transmit symbol represents
k log 2 M
PCM waveforms (line codes) are used for binary symbols (M=2).
10B
Energy per symbol in binary-PAM:
Binary PAM
(rectangular pulse)
4-ary PAM
(rectangular pulse)
3B
A.
T
T
10
0
-A.
-3B
T
-B
00
11
01
T
Nonreturn-to-zero (NRZ)
Return-to-zero (RZ)
NRZ-L
+V
-V
Phase encoded
Multilevel binary
+V
Manchester -V
Unipolar-RZ +V
0
Miller +V
-V
+V
Bipolar-RZ 0
-V
0
+V
Dicode NRZ 0
-V
2T 3T 4T 5T
0
2T 3T 4T 5T
Digital info.
Format
Textual
source info.
Analog
info.
Bit stream
(Data bits)
Sample
Sampling at rate
f s 1 / Ts
(sampling time=Ts)
Quantize
Encode
Pulse waveforms
(baseband signals)
Pulse
modulate
Rb mR
Receiver Structure
&
Matched Filter
estimated symbol
Format
m i
Detect
M-ary modulation
i 1, , M
n(t )
Demod.
z (T ) & sample r (t )
Impact of AWGN
hc (t ) (t ) 0.5 (t 0.75T )
Receiver job
Demodulation and sampling:
Waveform recovery and preparing the received signal
for detection:
Improving the signal power to the noise power
(SNR) using matched filter (project to signal space)
Reducing ISI using equalizer (remove channel
distortion)
Sampling the recovered waveform
Detection:
Estimate the transmitted symbol based on the
received sample
Receiver structure
Step 1 waveform to sample transformation
r (t )
Frequency
down-conversion
For bandpass signals
Received waveform
Receiving
filter
Detect
Equalizing
filter
z (T )
Threshold
comparison
Compensation for
channel induced ISI
Baseband pulse
(possibly distorted)
Baseband pulse
Sample
(test statistic)
m i
Baseband vs Bandpass
Bandpass model of detection process is equivalent to
baseband model because:
The received bandpass waveform is first transformed
to a baseband waveform.
Equivalence theorem:
Performing bandpass linear signal processing followed
by heterodying the signal to the baseband,
yields the same results as
heterodying the bandpass signal to the baseband ,
followed by a baseband linear signal processing.
Steps in design:
Model the received signal
Find separate solutions for each of the goals.
si (t )
r (t ) si (t ) h c (t ) n(t )
r (t )
hc (t )
n(t )
AWGN
Ideal channels
hc (t ) (t )
r (t )
si (t )
n(t )
AWGN
r (t ) si (t ) n(t )
h(t ) hopt (t ) si (T t )
*
H ( f ) H opt ( f ) S i ( f ) exp( j 2fT )
which is the time-reversed and delayed version of the conjugate of
the transmitted signal
h(t ) hopt (t )
si (t )
Correlator Receiver
r ( )si ( ) d r (t ), s (t )
*
Aside: Autocorrelation
Figure: shows how the signal can be located within the noise.
A copy of the known reference signal is correlated with the unknown
signal.
The correlation will be high when the reference is similar to the
unknown signal.
A large value for correlation shows the degree of confidence that the
reference signal is detected.
The large value of the correlation indicates when the reference
signal occurs.
Consider the received signal as a vector r, and the transmitted signal vector as s
Matched filter projects the r onto signal space spanned by s (matches it)
Filtered signal can now be safely sampled by the receiver at the correct sampling
instants, resulting in a correct interpretation of the binary message
Matched filter is the filter that maximizes the signal-to-noise ratio it can
be shown that it also minimizes the BER: it is a simple projection
operation
h opt (t )
A
T
A2
A
T
si (t )
h opt (t )
A
T
A
T
T/2 T
A
T
y (t ) si (t ) h opt (t )
2T
0 T/2 T 3T/2 2T
y (t ) si (t ) h opt (t )
A2
T/2 T
A
T
A2
2
The Fourier transform of a matched filter output with the matched signal as input
is, except for a time delay factor, proportional to the ESD of the input signal.
Z ( f ) | S ( f ) |2 exp( j 2fT )
2.
z (t ) Rs (t T ) z (T ) Rs (0) Es
3.
The output SNR of a matched filter depends only on the ratio of the signal energy
to the PSD of the white noise at the filter input.
S
max
N
4.
Es
N0 / 2
spectral phase matching that gives the desired output peak at time T.
spectral amplitude matching that gives optimum SNR to the peak value.
*
1
s (T t )
z1 (T )
z1
r (t )
*
sM (T t )
zM
z M
(T )
zi r (t ) s i (T t ) i 1,..., M
z ( z1 (T ), z 2 (T ),..., z M (T )) ( z1 , z 2 ,..., z M )
Note: we are projecting along the basis directions of the signal space
Implementation of correlator
receiver
Bank of M correlators
s 1 (t )
z1 (T )
r (t )
s
(t )
0
z1
Correlators output:
z Observation
vector
z M
z M (T )
z ( z1 (T ), z 2 (T ),..., z M (T )) ( z1 , z 2 ,..., z M )
T
zi r (t )si (t )dt
i 1,..., M
A
T
r (t )
z1 (T )
A
T
z1
s2 (t )
0
0
A
T
t
A
T
z2
z 2 (T )
SNR:
1 (t )
z ( z1 , z 2 )
s 3 (a31 , a32 )
s 2 (a21 , a22 )
s1 (t ) a11 1 (t ) a12 2 (t ) s1 (a11 , a12 )
Transmitted signal
s2 (t ) a21 1 (t ) a22 2 (t ) s 2 (a21 , a22 )
alternatives
Received signal at
matched filter output
Signal space
x(t ), y (t )
*
x
(
t
)
y
(t )dt
ax(t ), y (t ) a x(t ), y (t )
x(t ), ay (t ) a * x(t ), y (t )
x(t ) y (t ), z (t ) x(t ), z (t ) y (t ), z (t )
Signal space
What is norm?
Norm of a signal:
x(t ) dt E x
ax(t ) a x(t )
Norm between two signals:
d x , y x(t ) y (t )
s1 (a11 , a12 )
E1
d s1 , z
1 (t )
z ( z1 , z 2 )
E3
s 3 (a31 , a32 )
d s3 , z
E2
d s2 , z
s 2 (a21 , a22 )
The Euclidean distance between signals z(t) and s(t):
d si , z si (t ) z (t ) (ai1 z1 ) 2 (ai 2 z2 ) 2
i 1,2,3
j 1
i (t ), j (t ) i (t ) (t )dt K i ji
*
j
where
If all
0t T
j , i 1,..., N
1 i j
0i j
ij
Ki 1
2
cos(2t / T )
T
0t T
(t )
2
2
sin( 2t / T )
T
0t T
1 (t )
1 (t ), 2 (t ) 1 (t ) 2 (t )dt 0
2 (t )
1 (t )
1 (t ) 2 (t ) 1
Example:
1-dimensional orthonornal signal space
1 (t )
1
T
1 (t ) 1
0
T
1 (t )
Example: BPSK
Signal space
Any arbitrary finite set of waveforms si (t ) iM1
where each member of the set is of duration T, can
be expressed as a linear combination of N orthogonal
waveforms j (t ) Nj1 whereN M .
N
si (t ) aij j (t )
i 1,..., M
NM
j 1
where
T
1
1
aij
si (t ), j (t )
si (t ) *j (t )dt
Kj
Kj 0
j 1,..., N
0t T
i 1,..., M
N
Ei K j aij
j 1
Waveform energy
Signal space
N
si (t ) aij j (t )
j 1
1 (t )
ai1
si (t )
N (t )
aiN
1 (t )
ai1
aiN
sm
sm
ai1
aiN
ai1
N (t )
aiN
si (t )
s1 (a11 , a12 )
1 (t )
s 3 (a31 , a32 )
s 2 (a21 , a22 )
Transmitted signal
alternatives
aij si (t ) j (t )dt
0
j 1,..., N
i 1,..., M
0t T
z1
Observation
vector
r (t )
z1
N (T t )
zN
z N
si (t ) aij j (t )
j 1
i 1,..., M
NM
z ( z1 , z 2 ,..., z N )
z j r (t ) j (T t )
j 1,..., N
1 (t )
z1
r (t )
N (t )
si (t ) aij j (t )
zN
r1
rN
i 1,..., M
j 1
z ( z1 , z 2 ,..., z N )
NM
z j r (t ) j (t )dt
0
j 1,..., N
Observation
vector
1 (t )
s2 (t )
A
T
1
T
0
0
t
0
A
T
1 matched filter
1 (t )
r (t )
1
T
z1
z1 z
T t
n(t ) n (t ) n~ (t )
Noise projected on the signal space
(colored):impacts the detection process.
n (t ) n j j (t )
n (t )
Vector representation of
j 1
n j n(t ), j (t )
n~ (t ), (t ) 0
j
j 1,..., N
j 1,..., N
n (n1 , n2 ,..., nN )
independent zero-mean
Gaussain random variables
with variance var(n j ) N 0 / 2
j j 1
mi
Modulator
si
Decision rule
z si n
n
1
pn (n)
exp
N 0 N / 2 N 0
The elements of observed vector
z ( z1 , z 2 ,..., z N )
are independent Gaussian random variables. Its
2
pdf is
s
1
i
pz ( z | s i )
exp
N /2
N
N 0
0
Detection
Optimum decision rule (maximum a posteriori
probability):
Set m mi if
Pr(mi sent | z ) Pr(mk sent | z ), for all k i
where k 1,..., M .
Applying Bayes rule gives:
Set m mi if
pz (z | mk )
pk
, is maximum for all k i
pz ( z )
Detection
Z1 ,..., Z M
Vector z lies inside region Z i if
pz (z | mk )
ln[ pk
], is maximum for all k i.
pz ( z )
That means
m mi
Set m mi if
pz (z | mk ), is maximum for all k i
or equivalently:
Set m mi if
ln[ pz (z | mk )], is maximum for all k i
which is known as maximum likelihood.
Detection (ML)
Partition the signal space into M decision
regions,
Z1 ,..., Z M
Z2
s2
Z3
s3
s1
s4
Z4
Z1
1 (t )
p (z | m )dz
z
Zi
pz (z | m1 )
s2
Eb
s1
0
1 (t )
Eb
s1 s 2 / 2
Pe (m1 ) Pe (m2 ) Q
N 0 / 2
PB PE (2) Q
2 Eb
N0
mi )
PE ( M ) Pr ( m
i 1
1
PE ( M )
M
1
Pe (mi ) 1
M
i 1
1
1
M
P (m )
i 1
p (z | m )dz
i 1 Z i
Union bound
Union bound
Pe (mi ) P2 (s k , s i )
k 1
k i
1
PE ( M )
M
P (s
i 1 k 1
k i
, si )
p (r | m )dr
r
Z 2 Z3 Z 4
Z2
Z1
s2
s1
1
s3
Union bound:
s4
Z4
Z3
Pe (m1 ) P2 (s k , s1 )
k 2
A2 r
s2
s2
s1
s3
P2 (s 2 , s1 )
s4
p (r | m )dr
r
A2
s2
s1
2
s1
1
s3
P2 (s 3 , s1 )
A3
p (r | m )dr
r
A3
s4
1
s3
P2 (s 4 , s1 )
A4
s4
p (r | m )dr
r
A4
d ik / 2
1
u2
exp( )du Q
N /2
N0
N 0
0
d ik
d ik s i s k
1
PE ( M )
M
d min / 2
P2 (s k , s i ) ( M 1)Q
N /2
i 1 k 1
0
k i
d min min d ik
i ,k
ik
s i Ei Es , i 1,...,4
d i ,k 2 Es
ik
Es
d min 2 Es
s3
Es
s2
d1, 2
d 2,3
s1
d 3, 4
d1, 4
s4
Es
Es
1 (t )
Eb
STb
S W
N 0 N / W N Rb
Rb
W
: Bit rate
: Bandwidth
Binary PAM
s2
s1
Eb
s4
6
Eb
5
1 (t )
Eb
4-ary PAM
s3
s2
E
2 b
5
E
2 b
5
s1
6
Eb
5
1 (t )
1
T
T t
1 (t )
MPAM:
Error probability:
Complex
Detection
:
Summary
Pe is proportional to
BER
Average SNR
Diversity of
L:th order
( Pe )
AWGN
channel
(no fading)
SNR
L=4
L=3
L=2
( 0)
Modulation Techniques
What is Modulation?
How?
Digital vs Analog
Modulation
Modulation: representation
Any modulated signal can be represented as
s(t) = A(t) cos [ ct + (t)]
amplitude
phase or frequency
MPSK
Circular
Square
I(t), in-phase
Q(t), quadrature
LINEAR MODULATIONS
Square
Constellations
M-ARY QUADRATURE
AMPLITUDE MOD.
(MQAM)
M 4
M-ARY PHASE
SHIFT KEYING
(MPSK)
M=4
(4-QAM = 4-PSK)
CONVENTIONAL
4-PSK
(QPSK)
Circular
Constellations
M 4
OFFSET
DIFFERENTIAL
4-PSK
4-PSK
(OQPSK) (DQPSK, /4-DQPSK)
4-PSK
16-PSK
an
an
Tradeoffs
Higher-order modulations (M large) are more spectrally
efficient but less power efficient (i.e. BER higher).
M-QAM is more spectrally efficient than M-PSK but
also more sensitive to system nonlinearities.
MQAM:
MFSK:
Source: Rappaport book, chap 6
Binary PAM
s2
s1
Eb
Eb
5
Eb
4-ary PAM
s3
s2
s4
6
1 (t )
Eb
5
Eb
5
s1
6
Eb
5
1 (t )
Gray coding
00
s4
6
Eb
5
01
11
10
4-ary PAM
s3
s2
E
2 b
5
E
2 b
5
s1
6
Eb
5
1 (t )
MPAM: Details
Unequal energies/symbol:
Decision Regions
MPSK:
4-PSK
16-PSK
an
Constellation points:
Gray coding
01
01
00
00
11
10
11
10
Z3
Z4
Z1
Z2
Z5
Z1
Z6
Z4
4PSK
Z7
Z8
8PSK
si(t) + n(t)
g(Tb - t)
Z1 : r > 0
Z2 : r 0
m=0
cos(2fct)
m = 0 or 1
MQAM:
Z1
Z2
Z3
Z4
Z5
Z6
Z7
Z8
Z9
Z10
Z11
Z12
Z13
Z14
Z15
Z16
Differential Modulation
(Contd)
Quadrature Offset
s(t) = A cos [ ct + 2 kf
d( ) d]
nonlinear modulation but constant-envelope
Spectral Characteristics
10
0
QPSK/DQPSK
GMSK
-20
-40
-60
B3-dBTb = 0.16
0.25
-80
1.0
-100
(MSK)
-120
0
0.5
1.0
1.5
2.0
2.5
For Pb = 10-3
10-3
5
BPSK, QPSK
BPSK 6.5 dB
QPSK 6.5 dB
DBPSK ~8 dB
DQPSK ~9 dB
DBPSK
DQPSK
10-4
5
dB power
penalty for
differential
detection.
-5
5
2
10-6
There is ~3
2
10
10
b, SNR/bit, dB
12
14
2
10-1
5
2
10-2
5
Pb
DBPSK
2
BPSK
10-3
5
AWGN
2
10-4
5
2
10-5
10
15
20
b, SNR/bit, dB
25
30
35
100
QPSK
DQPSK
10 -1
Rayleigh Fading
10 -2
For fD = 80 Hz,
data rate
-3
Pb 10
fDT=0.003
No Fading
10 -4
0.002
0.001
10 -5
0
10 -6
10
20
30
40
b, SNR/bit, dB
50
60
Pbfloor
10 kbps
10-4s
3x10-4
100 kbps
10-5s
3x10-6
1 Mbps
10-6s
3x10-8
The implication is that Doppler is not an issue for high-speed
wireless data.
[M. D. Yacoub, Foundations of Mobile Radio Engineering ,
CRC Press, 1993]
25 sec
2.5 sec
500 nsec
100 nsec
8 kbps
80 kbps
400 kbps
2 Mbps
[J. C.-I. Chuang, "The Effects of Time Delay Spread on Portable Radio
Communications Channels with Digital Modulation,"
IEEE JSAC, June 1987]
Coherent Detection
+ BPSK
QPSK
OQPSK Modulation
x MSK
10-2
x
x
+
+
x
+
10-3
+
x
+
10-4
10-2
10-1
rms delay spread
=
symbol period
T
100
Limitations
flat fading
doppler
delay spread
Pulse Shaping
hc (t ) (t ) 0.5 (t 0.75T )
Hc ( f ) Hc ( f ) e
j c ( f )
Non-constant amplitude
Non-linear phase
Amplitude distortion
Phase distortion
H ( f ) Ht ( f )H c ( f )H r ( f )
ISI effect
z k sk nk i si
ik
x1 x2
Channel
Tx filter
x3
Ht ( f )
xk
x1 x2
T
hr (t )
Hc ( f )
Hr ( f )
zk
t kT
Detector
xk
n(t )
Equivalent model
Equivalent system
x3
r (t ) Rx. filter
h(t )
H( f )
zk
z (t )
t kT
H ( f ) Ht ( f )H c ( f )H r ( f )
n (t )
filtered noise
Detector
xk
Rs
Rs
1
W
2 [symbol/s/Hz]
2T
2
W
Bandwidth efficiency, R/W [bits/s/Hz] :
An important measure in DCs representing data
throughput per hertz of bandwidth.
Showing how efficiently the bandwidth resources are
used by signaling techniques.
H( f )
h(t ) sinc(t / T )
1
2T
1
2T
1
2T
2T T
T 2T
Reduce ISI
| H ( f ) || H RC ( f ) |
r 0
r 0.5
0.5
1 3 1
T 4T 2T
r 1
1 3
2T 4T
Rs
Baseband W sSB (1 r )
2
1
T
r 1
0.5
3T 2T T
r 0.5
r 0
2T
Passband W DSB (1 r ) Rs
3T
Raised-Cosine Filter
| f | W 2W0
H ( f ) cos
4
W
W
0
0
2
for | f | 2W0 W
for 2W0 W | f | W
for | f | W
cos[2 (W W0 )t ]
h(t ) 2W0 (sinc(2W0t ))
1 [4(W W0 )t ]2
Excess bandwidth: W W0
Roll-off factor r
0 r 1
W W0
W0
H RC ( f ) H t ( f ) H c ( f ) H r ( f ) H e ( f )
Square-Root Raised Cosine (SRRC) filter and Equalizer
H RC ( f ) H t ( f ) H r ( f )
H r ( f ) H t ( f ) H RC ( f ) H SRRC ( f )
He ( f )
1
Hc ( f )
t/T
First pulse
Second pulse
Data symbol
Amp. [V]
Raised Cosine pulse at theBaseband
output of
matched
received
waveform at
filter
the matched filter output
(zero ISI)
t/T
Eye pattern
amplitude scale
Sensitivity to
timing error
Timing jitter
time scale
Summary
Digital Basics
Pulse Shaping
Extra Slides
Bandpass Modulation: I, Q
Representation
Fading improvement using in-band pilot tones & adapt receiver gain
to compensate