Professional Documents
Culture Documents
Telecommunications
Engineering II
Jorma Kekalainen
Contents
Topics include
Signals and noise
Fourier analysis
Digital transmission
Statistical and linear algebra tools for advanced
telecommunications
Multi-carrier modulation
Coding and information theory
Diversity techniques
Wireless overview
1
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Information Sources
1. Usually, the most precise sources are the original
sources, i.e. standards, recommendations or other
specifications.
You can pull them from the Internet e.g.
- ITU-T www.itu.int/ITU-T/
- IETF www.ietf.org
- 3GPP www.3gpp.org
or from elsewhere.
Books
Carlson et al.: Communication Systems: An Introduction to Signals and
Noise in Electrical Communication
Haykin: Communication Systems
Haykin & van Veen: Signals & Systems
Freeman: Radio System Design for Telecommunications
Goldsmith: Wireless Communications
Murthy et al.: Ad Hoc Wireless Networks: Architectures and Protocols
Pahlavan et al.: Principles of Wireless Networks: A Unified Approach
Rappaport: Wireless Communications: Principles and Practice
Roddy: Satellite Communications
Skolnik: Introduction to Radar Systems
Stallings: Wireless Communication and Networks
Tse: Fundamentals of Wireless Communication
2
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Introduction
3
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Measure of information
4
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Transmission channels
Cables
wire pairs (e.g., ordinary telephone line)
coaxial cable
waveguide (metallic waveguide and optical fiber)
More or less free space radio transmission
broadcasting
point-to-point microwave transmission
satellite position transmission
cell networks
(Portable magnetic/electronic/optical memory equipment)
Mode of communication
5
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
6
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Constraints
Reliability of communication
7
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Noise
Transmitted Received Received
Info. signal signal info.
SOURCE
Source Transmitter Channel Receiver User
Transmitter
Source Channel
Formatter Modulator
encoder encoder
Receiver
Source Channel
Formatter Demodulator
decoder decoder
8
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Textual Format
source info.
Pulse
Analog Transmit
Sample Quantize Encode modulate
info.
Pulse
Bit stream waveforms Channel
Format
Analog
info. Low-pass
Decode Demodulate/
filter Receive
Textual Detect
sink
info.
Digital info.
9
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
10
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Time functions
21
22
11
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
23
24
12
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
25
26
13
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Continuous-time signals
27
Sine signal
28
14
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Sinusoids
29
Cosine signal
30
15
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Equivalence of sinusoidals
31
32
16
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Exponentials
33
Other classification
34
17
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Period
35
Complex exponentials
36
18
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Complex exponential
37
Complex exponential
38
19
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
39
40
20
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
41
Energy signals
42
21
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Power signals
43
Impulse function
22
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Impulse
45
Dirac-delta function
46
23
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Application
47
Advanced definition
48
24
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Step function
49
Another definition
50
25
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Step function
51
Application
52
26
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
53
Synthesis
54
27
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Signum function
55
Rectangular pulse
56
28
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Rectangular pulse
Also notation:
57
58
29
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Amplitude scaling
59
Examples
60
30
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Time scaling
61
Note: Downsampling is sometimes called decimation
Examples
62
31
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Time shifting
63
Examples
64
32
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Reflection
65
LTI systems
66
33
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
System
67
68
34
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
69
System: Definition
70
35
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
71
72
36
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
73
74
37
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
75
a1 aN are any arbitrary constants, and x1(t) xN(t) are any arbitrary continuous-
time signals.
a1 aN are any arbitrary constants, and x1(n) xN(n) are any arbitrary discrete-
time signals
76
38
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
LTI system
77
Linearity condition
78
39
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Linearity condition
79
Note
80
40
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Principle of superposition
81
Time-invariance
82
41
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Impulse response
83
84
42
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Impulse response
85
Example
86
43
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
87
Convolution
88
44
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Convolution
89
Convolution
90
45
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Impulse response
91
Causal system
92
46
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
93
94
47
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
95
96
48
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
97
Examples: Interconnections of
systems/components
Series: e.g. transmitter-channel-receiver
98
49
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
99
Step response
100
50
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
101
102
51
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
103
104
52
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Fourier Analysis
106
53
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
107
Complex exponential as an
eigenfunction
108
Note: Only one frequency, namely f0.
54
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
110
55
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Frequency content
111
Fourier transform
112
56
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note:
113
114
57
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
115
116
58
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Spectrum of sinusoid
117
118
59
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
119
120
60
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
121
122
61
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note: 123
124
62
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
125
126
63
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
127
F-representations
128
64
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Existence of F-representations
129
Gibbs effect
130
65
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Gibbs effect
131
Properties of F-representations
132
66
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Properties of F-representations
133
Properties of F-representations
134
67
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note
135
136
68
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
137
138
69
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
139
F{convolution}
140
70
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
y=xh XH=Y
141
142
71
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Energy
143
Parsevals formula
144
72
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Power spectrum
145
146
73
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
147
148
74
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
149
150
75
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Modulation
151
Modulation
152
76
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Periodic extension
153
154
77
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note: The DTFS is also called the Discrete Fourier Transform (DFT)
155
156
78
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
157
158
79
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
159
(y=xh XH=Y)
160
80
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Frequency response
161
81
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Distortionless system
163
Distortionless system
164
82
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
165
Ideal filters
166
83
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Ideal filters
Ideal filters are physically unrealizable, in the sense that their
characteristics cannot be achieved with a finite number of
elements.
The filters used in practical real time applications must be causal, that is
Realizable filters
168
84
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
169
Equalizers
If we know the frequency response of a system then, given the
output, we can discover the input using the equation
H(j)0
Example: Loading coils on telephone lines. Inductors are placed in shunt across the line
every km or so to improve frequency flatness over voice frequencies.
85
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: Equalization
If the transfer function of the channel that causes linear
distortion is known, it can be compensated by the inverse
transfer function.
H(f) = Hc(f)Heq(f)
Channel Equalizer
171
172
86
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
173
87
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
175
88
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
177
178
89
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note:
180
90
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note:
181
182
Spectral density function G(f) represents the distribution of the power or energy in the
frequency domain. The area under G(f) equals the average power or total energy.
91
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
183
92
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
185
186
93
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Inner product
The inner product for two m 1 complex vectors
is given by
The complex conjugate of a vector or row vector x, denoted as x*, is obtained by taking the
complex conjugate of each element of x. The Hermitian of a vector x, denoted as xH, is187
its
conjugate transpose: xH = (x)T.
188
94
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
189
190
95
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Autocorrelation function
The inverse Fourier transform of the energy spectral density Es(f) is
termed the autocorrelation function Rs(), since it measures how closely
the signal s matches delayed versions of itself.
Since |S(f)|2 =S(f)S(f)=S(f)SMF(f), where sMF(t)=s(t) is the matched
filter for s introduced earlier.
We therefore have that
Thus, Rs() is the outcome of passing the signal s through its matched
filter, and sampling the output at time , or equivalently, correlating the
signal s with a complex conjugated version of itself, delayed by .
While the preceding definitions are for finite energy deterministic
signals, we revisit these concepts in the context of finite power random
processes later.
191
192
96
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Real signals
193
Real signals
194
97
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Conjugate symmetry
Since bandpass signals are real, their frequency response has
conjugate symmetry, i.e. a bandpass signal s(t) has
for some f B.
This asymmetry in |S(f)| is illustrated in the previous figure.
Bandpass signals result from modulation of a baseband signal by
a carrier, or from filtering a deterministic or random signal with
a bandpass filter.
195
Bandpass signal
196
98
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Bandpass signal
197
Canonical representation
198
99
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Direct-conversion modem
199
Polar decomposition
200
100
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
201
A baseband communication refers to a system that does not include modulation.
202
101
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
where fc>W>0.
A channel modeled as a linear time-invariant system is said to be
passband if its transfer function H(f) satisfies the previous
equation
203
The spectrum S(f) for a real- The spectrum S(f) for a real-valued passband
valued baseband signal. The signal. The bandwidth of the signal is B.
bandwidth of the signal is B. The figure shows a frequency fc within the
band in which S(f) is nonzero. Typically, fc is
much larger than the signal bandwidth B.204
102
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
103
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
I- and Q-components
The waveforms sc(t) and ss(t) are also referred to as the in-phase (or I)
component and the quadrature (or Q) component of the passband signal
sp(t), respectively.
(2)
(3)
Note: To check (3), plug in (2) and Eulers identity on the right-hand side (3) 208
to obtain the expression (1).
104
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
(4)
Plugging
into (3), we obtain yet another formula for the passband signal s:
(5)
The equations (1), (3) and (5) are three different ways of expressing
the same relationship between passband and complex baseband in the
time domain.
209
That is,
210
105
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
211
212
106
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Noise
Transmitted Received Received
Info. signal signal info.
SOURCE
Source Transmitter Channel Receiver User
Transmitter
Source Channel
Formatter Modulator
encoder encoder
Receiver
Source Channel
Formatter Demodulator
decoder decoder
107
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Textual Format
source info.
Pulse
Analog Transmit
Sample Quantize Encode modulate
info.
Pulse
Bit stream waveforms Channel
Format
Analog
info. Low-pass
Decode Demodulate/
filter Receive
Textual Detect
sink
info.
Digital info.
Introduction
216
108
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
217
Sampled signals
218
109
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
219
220
110
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
221
222
111
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
223
224
112
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
225
A paralleltoserial (P/S) converter
Sampling concepts
Continuous-time signal
113
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
227
Impulse train
228
114
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Periodic sampling
229
230
115
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
231
232
116
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Sampling theorem
233
Reconstruction as interpolation
234
117
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Reconstruction as interpolation
235
Bandlimited interpolation
236
118
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
237
238
119
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
239
240
120
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
241
242
121
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
243
Bandwidth of signal
Baseband versus bandpass:
Baseband Bandpass
signal signal
Local oscillator
Bandwidth dilemma:
Bandlimited signals are not realizable!
Realizable signals have infinite bandwidth!
244
122
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
(a)
(b)
(c)
(d)
245
(e)50dB
x (t ) | X ( f ) |
xs (t )
| Xs( f )|
246
123
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
LP filter
Nyquist rate
aliasing
247
248
124
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Sampling theorem
Analog Sampling Pulse amplitude
signal process modulated (PAM) signal
In practice, it is need to sample faster than this because the receiving filter
will not be sharp.
249
Sampling theorem
Statement: Any signal with a bandwidth of W can be completely reconstructed if it is
sampled at a rate of 2W.
Original
waveform
samples
Capacitor
discharges
Capacitor
charges
Thus by sampling first at the transmitter and then passing the samples through a ideal
LPF the original waveform can be completely reconstructed 250
125
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Original Incorrectly
signal at 3 Hz Samples at less reconstructed
sampling rate at 2 Hz signal at 1 Hz
Thus when any wave is sampled at a frequency that is less than double
the maximum signal frequency, the recovered wave will not be of the same
frequency as the input waveform. This distortion is called aliasing .
The sampling frequency has to be adjusted such that fs > 2fm 251
Pulse modulation
252
126
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Pulse modulation
Signal
PAM
PWM
PPM
253
Encoding (PCM)
254
127
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
256
A paralleltoserial (P/S) converter
128
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
257
129
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
260
130
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
261
quantizer digitized
sampling voice
and
circuit signal
compander
analog
voice voice band
signal filter
262
131
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Quantization
263
Memoryless quantization
264
132
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Uniform quantizer
265
266
133
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Uniform quantization
Amplitude quantizing: Mapping samples of a continuous
amplitude waveform to a finite set of amplitudes.
Out
In
Average quantization noise power
Quantized
267
Example
Derive quantization noise for uniform quantization in case of
signal x[-1,1] and the number of quantization levels is M.
Quantization error is
134
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
0-level
269
Example
If Max{j}<< then
In whole range
270
135
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
Note:
So
271
Example
If j= = constant j, then
136
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: SNRqdB
Derive SNRqdB for uniform quantization in case of uniformly distributed
and normalized signal |x(t)|1 and the number of quantization levels is M.
273
Example: SNRqdB
So
and
137
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
275
Quantization error
Quantizing error: The difference between the input and
output of a quantizer e(t ) = x (t ) x (t )
+
e(t ) =
x (t ) x (t )
276
138
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Non-linear quantization
277
278
139
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
279
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 companding
y = C (x) x
x(t ) y (t ) y (t ) x (t )
x y
Compress Quantize Expand
Transmitter Channel Receiver
280
140
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Non-linear quantization
The voltage range between the lowest level and the
highest level is divided into segments in a non-linear
manner logarithmic
281
Companding
During the companding process, input analog signal
samples are compressed into logarithmic segments and
then each segment is quantized and coded using uniform
quantization.
282
141
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Speech compression
=255
Log.
Lin.
A=87.6
283
284
142
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
285
CODE 1101
1.2
Vi
1.2
CODE 0010
CODE 0000
286
expanding
143
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Differential quantizier
287
1-point DPCM
288
144
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Baseband transmission
To transmit information through 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 =log2 M bits of the PCM words.
PCM waveforms (line codes) are used for binary symbols (M=2).
289
290
145
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
3B
A.
11
1 B
T
T T 01
T -B 00 T T
0 10
-A. -3B
291
Nonreturn-to-zero (NRZ)
Return-to-zero (RZ)
+V 1 0 1 1 0 +V 1 0 1 1 0
NRZ -V Manchester -V
Unipolar-RZ +V Miller +V
0 -V
+V +V
Bipolar-RZ 0 Dicode NRZ 0
-V -V
0 T 2T 3T 4T 5T 0 T 2T 3T 4T 5T
292
146
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
293
147
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Codec
sampling
quantizing
encoding
Analog Digital
signal signal
Sampler Quantizer Encoder
295
M U
Analog Digital O P
signal C signal D C
O U O RF signal
D TDM L N
E Digital A V
C Base Band T E
O R
R T
296
148
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
297
298
149
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
299
300
150
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Reconstruction of m(t)
301
Bandlimited interpolation
302
151
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Sampling theorem
A signal having no frequency components above W Hertz is
completely described by specifying the values of the signal at
periodic time instants that are separated by at most 1/(2W)
seconds.
303
Natural sampling
304
152
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
305
306
153
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Flat-top sampling
307
308
154
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Equalization
309
Pulse modulation
310
155
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
311
Quantization
312
156
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Memoryless quantization
313
Uniform quantizer
314
157
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
315
316
158
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
317
318
159
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Optimal quantizer
319
Optimal quantizer
320
160
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
321
322
161
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Robust quantizers
323
324
162
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
325
326
163
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
327
328
164
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
One sacrifices performance for larger input power levels to obtain a 329
performance that remains robust over a wide range of input levels.
330
165
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Differential quantizers
331
Linear predictor
332
166
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
333
334
167
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
335
Intersymbol Interference
(ISI)
Pulse shaping and
equalization
336
168
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
337
338
169
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
339
340
170
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Bandwith constraint
341
342
171
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
343
Received
344
172
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
345
346
173
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
347
ISI
This spreading and smearing of symbols such that the energy
from one symbol effects the next ones in such a way that the
received signal has a higher probability of being interpreted
incorrectly is called Inter-Symbol-Interference or ISI.
ISI can be caused by many different reasons.
It can be caused by filtering effects from hardware or frequency
selective fading, and from non-linearity effects.
Communication system designs for both wired and wireless
nearly always need to incorporate some way of controlling ISI.
348
174
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
H c ( f ) = H c ( f ) e j c ( f )
H ( f ) = Ht ( f )H c ( f )H r ( f )
creates echoes and hence time dispersion
causes ISI at sampling time
ISI effect
z k = sk + nk + i si
ik
350
175
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
x1 x2
{xk } Equivalent system
h(t )
z (t ) zk
{xk }
Detector
t = kT
T H( f )
x3 T n (t )
filtered noise
H ( f ) = Ht ( f )H c ( f )H r ( f )
351
352
176
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Pulse shaping
353
354
177
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
355
356
178
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Pulse shaping
The main tool used to counter ISI is pulse shaping.
How can pulse shaping help control ISI?
The secret lies in the digital demodulation process used.
When the timing pulse slices the signal to determine the value
of the signal at that instant, it does not care what the signal
looked like before or after it.
So if there was some way we could keep the symbols from
interfering in such a way that they do not affect the amplitude
at the slicing instant, we can counter ISI successfully.
357
179
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1 R R
= s W s 2 [symbol/s/Hz]
2T 2 W
Bandwidth efficiency, R/W [bits/s/Hz] :
An important measure in DCs representing data
throughput per Hz of bandwidth.
Showing how efficiently the bandwidth resources are used
by signaling techniques. 359
0 f 2T T 0 T 2T t
1 1
2T 2T
1
W= 360
2T
180
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
362
181
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
The frequency
response of the
square pulse is a
sinc function.
Lowpass
bandwidth is one
half of the
bandpass case.
363
364
182
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
365
Pulse bandwidth
366
183
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
An important relationship
367
368
184
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Duality
369
Example
370
185
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
371
Nyquist bandwidth
The bandwidth achieved by the sinc pulse is called the Nyquist
bandwidth.
It requires only 1/2 Hz per symbol.
Can we find something even better?
It turns out that we have not been able to find any other shape
that can improve on this.
It is an ultimate limit for perfect reconstruction of the signal.
Band-limited spectrum in frequency domain with no energy
going to waste and small total bandwidth requirement seems to
be great!
But not so great however, because a sinc pulse is actually no
more possible to build than is a square pulse.
372
186
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
373
187
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
375
r = 0.5
0.5 0.5 r =1
r =1 r = 0.5
r =0
1 3 1 0 1 3 1 3T 2T T 0 T 2T 3T
T 4T 2T 2T 4T T
Rs
Baseband W sSB= (1 + r ) Passband W DSB= (1 + r ) Rs
2
376
r = roll-off factor
188
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Raised cosine
Nyquist offered ways to build (realizable) shapes that had the
same good qualities as the sinc pulse and less of the
disadvantages.
One class of pulses he proposed are called the raised cosine
pulses.
They are really a modification of the sinc pulse.
The sinc pulse has a bandwidth of W0, where W0 is specified
as
W0 = 1/2T
The raised cosine pulses have an adjustable bandwidth which
can be varied from W0 to 2W0.
We want to get as close to W0, which is called the Nyquist
bandwidth.
377
Roll-off factor
The factor r related the achieved bandwidth to the ideal bandwidth W0 as
W W0
Roll-off factor r =
W0
0 r 1
where W0 is Nyquist bandwidth, and W is the utilized bandwidth.
The factor r is called the roll-off factor.
It indicates how much bandwidth is being used over the ideal bandwidth.
The smaller this factor, the smaller bandwidth and the more efficient the
scheme.
The percentage over the minimum required W is called the excess
bandwidth.
It is 100% for roll-off of 1.0 and 50% for roll-off of 0.5.
378
189
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Roll-off
379
380
190
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Roll-off
factor r=
381
Roll-off
factor r=
382
191
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
t/T
383
Root-raised cosine
The whole raised cosine can be applied at once at the transmitter but in
practice it has been found that concatenating two filters each with a root
raised cosine response (called split-filtering) works better.
So to implement the raised cosine response, we split the filtering in two
parts to create a matched set.
In frequency domain, we take the square root of the frequency response
hence the name root-raised cosine.
Split filtering of raised cosine response, a root-raised cosine filter at the transmitter and
one at the receiver, giving a total response of a raised cosine.
384
192
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
385
386
193
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
387
Third pulse
t/T
First pulse
Second pulse
Data symbol
388
194
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
389
Eye diagram
195
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Eye pattern
Eye pattern: Display on an oscilloscope which sweeps
the system response to a baseband signal at the rate
1/T (T symbol duration) the superposition of
successive symbol intervals
Distortion
due to ISI
Noise margin
amplitude scale
Sensitivity to
timing error
Timing jitter
time scale 391
392
196
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
393
394
197
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
395
396
198
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
397
path-1
path-2
path-3
multi-path path-2
propagation
Path Delay
path-1
path-3
Mobile Station (MS)
Base Station (BS)
398
199
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
399
400
200
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Multipath
effects
~ O(1s)
401
Set of multipaths
changes ~ O(5 ms)
402
201
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
403
Note 1: At each tap, channel gain |h| is a Rayleigh distributed r.v.. The
random process is not IID.
Note 2: Response spreads out in the time-domain (), leading to inter-
symbol interference and deep fades in the frequency domain:
frequency-selectivity caused by multi-path fading
Note 3: Response completely vanish (deep fade) for certain values of t:
Time-selectivity caused by doppler effects (frequency-domain 404
dispersion/spreading)
202
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Dispersion-selectivity duality I
405
Dispersion-selectivity duality II
406
203
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Fading terminology
Flat fading: no multipath ISI effects.
E.g. narrowband, indoors
Frequency-selective fading: multipath ISI effects.
E.g. broadband, outdoor.
Received Signals:
Line-of-sight:
Reflected:
204
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
What is an equalizer?
409
Equalization
z (T ) Threshold m i
r (t ) Frequency Receiving Equalizing
comparison
down-conversion filter filter
410
205
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
H c ( f ) = H c ( f ) e j c ( f )
H RC ( f ) = H t ( f ) H c ( f ) H r ( f ) H e ( f )
1
He ( f ) = Taking care of ISI
Hc ( f ) caused by channel
206
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
413
414
207
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Equalizing filters
Baseband system model
a1
a (t kT ) Tx filter
k Channel r (t ) Equalizer Rx. filter z (t ) z k {a k }
k ht (t ) hc (t ) he (t ) hr (t ) Detector
t = kT
Ta a Ht ( f ) Hc ( f ) He ( f ) Hr ( f )
2 3
n(t )
Equivalent model H ( f ) = H t ( f )H c ( f )H r ( f )
a1
a (t kT )
k
Equivalent system z (t ) x(t ) Equalizer z (t )
zk {ak }
k h(t ) he (t ) Detector
t = kT
Ta a H( f ) He ( f )
2 3 n (t )
filtered (colored) noise
n (t ) = n(t ) hr (t )
415
Equalizer types
416
208
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Filter coefficients
The idea behind RLS filters is to minimize a cost function C by
appropriately selecting the filter coefficients wn , updating the
filter as new data arrives.
The error signal e(n) and desired signal d(n) are defined in the
negative feedback diagram below:
209
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Cost function
The weighted least squares error function C the cost function
we desire to minimize being a function of e(n) is therefore
also dependent on the filter coefficients:
419
Linear equalizer
A linear equalizer effectively inverts the channel.
n(t)
Equalizer
Channel
1
Hc(f) Heq(f)
Hc(f)
210
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
421
Feedback
Filter
211
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
x (t )
c N cN +1 c N 1 cN
z (t )
Coefficient
adjustment 423
424
212
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Adjust 1 k =0
z (k ) =
{cn }nN= N 0 k = 1,..., N
Adjust
{c n }nN= N
[
min E ( z (kT ) ak ) 2 ]
425
Equalizer
The ideal equalizer, an exact inverse system for the channel, is almost
always unrealizable.
With enough prior information about the channel, very good
approximations can be realized.
Using filter theory, the combination of channel and equalizer can be
made a near distortionless system.
Less distortion higher order filter, more delay.
Equalization can be done at receiver, transmitter or both.
However, the channel often is not known precisely at the time of
system design and needs to be fine-tuned during operation.
One approach to simplify the design of the equalizer is to focus on
eliminating ISI, rather than a complete inverse system.
At the input to the decision device, we simply try to enforce the zero-
ISI condition
peq(t0) = 1 and peq(t0 + nD) = 0 for n = 1, 2, ....
213
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Zero-forcing equalizer
Consider a transversal-filter equalizer positioned between the receiver
filter and the decision device.
427
Zero-forcing equalizer
When the system commences operation, the values of pR(t0 - kD)
for k = -2N, ... , 2N are measured during a training phase.
The system of equations (1) is then a set of 2N+ 1 linear equations
in 2N + 1 unknowns: the tap gains c[n] easily solved.
Typically the equalizer (and the solution of the linear equa-
tions) is implemented digitally.
This approach attempts to zero out as much of the ISI as possible:
hence called a zero-forcing equalizer.
A side effect may be noise enhancement: the noise power input
to the decision device may increase.
In many scenarios, the channel response may vary with time.
In this case, we may need to periodically suspend transmission
of data while the equalizer is re-trained.
More advanced equalizers are able to update continuously or
428
track the channel.
This is called adaptive equalization.
214
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Frequency-selective channel
(no equalization)
Gaussian
channel Flat fading channel
(no fading)
S/N
429
Frequency-selective channel
(no equalization)
Gaussian
channel Flat fading channel
(no fading)
S/N
430
215
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Frequency-selective channel
(no equalization)
Gaussian
channel Flat fading channel
(no fading)
S/N
431
S/N
432
216
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Equalization: Summary
Equalizer equalizes the channel response in frequency domain to
remove ISI
Can be difficult to design/implement,
Can get noise enhancement (linear EQs) or error propagation
(decision feedback EQs)
434
217
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Introduction
218
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
219
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
220
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Union Bound
A B
Applications:
Getting bounds on BER (bit-error rates),
Bounding the tails of probability distributions
Classic Experiment:
Tossing a die:S = {1,2,3,4,5,6}
Any subset A of S is an event:
A = {the outcome is even} = {2,4,6}
221
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
ii.) P(S) = 1
A B
P ( A B ) = P ( A ) + P (B )
A B =
Probability of events
In fact for any sequence of pair-wise-mutually-
exclusive events, we have
Ai A j = , and UA i =
S.
i =1
A1
A2
Ai P An = P ( An )
A
n =1 n =1
j An
222
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Conditional probability
223
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note that: P ( AB ) = P (B )P ( A | B ) = P ( A )P (B | A )
Also: P ( A | B ) = P ( A) and P (B | A ) = P (B )
Random variables
224
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
225
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
FX ( x) = P( X x) = P({s S | X ( s ) x})
Note that FX ( x ) is non-decreasing in x, i.e.
x1 x2 Fx ( x1 ) Fx ( x2 )
Also lim Fx ( x) = 0 and lim Fx ( x) = 1
x x
Plots of cdf
226
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
0 .9 L o g n o rm a l(0 ,1 )
G a m m a (.5 3 ,3 )
0 .8 E x p o n e n tia l(1 .6 )
W e ib u ll(.7 ,.9 )
0 .7 P a re to (1 ,1 .5 )
0 .6
F(x)
0 .5
0 .4
0 .3
0 .2
median
0 .1
0
0 2 4 6 8 10 12 14 16 18 20
x
-1
10
-2
10
log(1-F(x))
-3
10 L o g n o rm a l(0 ,1 )
G a m m a (.5 3 ,3 )
E x p o n e n tia l(1 .6 )
W e ib u ll(.7 ,.9 )
-4
10 P a re to II(1 ,1 .5 )
P a re to I(0 .1 ,1 .5 )
-1 0 1 2
10 10 10 10
lo g (x )
227
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Class Freq.
Count 15 but < 25 3
5 25 but < 35 5
35 but < 45 2
Frequency 4
Relative 3
frequency Bars
2
Percent
1
0 15 25 35 45 55
Lower Boundary
228
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
f (x )dx = 1
x
All X a b
(Area Under Curve)
f ( x ) 0, a x b Value
229
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
230
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
231
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1
f(x)
0 .5
0
0 0 .5 1 1 .5 2 2 .5 3 3 .5 4 4 .5 5
x
232
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Central Tendency
(Location)
Variation (Dispersion)
Shape
233
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Numerical data:
Properties & measures
Numerical Data
Properties
Central
Variation Shape
Tendency
Mean Range Skew
Median Inter-quartile Range
Mode Variance
Standard Deviation
234
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Expectation
The expectation (average) of a continuous random variable X
is given by
E( X ) = xf
X ( x)dx
235
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
236
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Indices/Measures of spread/dispersion
Why care?
237
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
238
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
239
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
240
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Correlation coefficient
241
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Characteristic function
The distribution of a random variable X can be determined from its
characteristic function, defined as
This will become significant in finding the distribution for sums of random
variables.
242
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
PX (1) = P( X = 1) = p PX (0) = P( X = 0) = 1 p
A discrete random variable that takes two values 1 and 0 with
probabilities p and 1-p.
Good model for a binary data source whose output is 1 or 0.
Can also be used to model the channel errors.
243
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Binomial distribution
P(X)
.6
.4 n = 5 p = 0.1
.2
.0 X
Mean 0 1 2 3 4 5
= E ( x ) = np
P(X) n = 5 p = 0.5
Standard Deviation .6
.4
.2
= np (1 p) .0 X
0 1 2 3 4 5
244
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Binomial distribution
Binomial can looks like skewed or normal
Depends upon
p and n !
25.00%
30.00%
25.00% 20.00%
20.00%
Number of Blocks
Number of Blocks
15.00%
15.00%
10.00%
10.00%
5.00%
5.00%
0.00% 0.00%
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Num ber of Losses out of N = 20 Number of Losses out of N = 20
16.00%
distribution near the mean: 14.00%
12.00%
8.00%
50% PER
4.00%
Npq = 5
2.00%
0.00%
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Num ber of Losses out of N = 20
245
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
x
P( X = x) = e x = 0,1, 2,...
x!
where > 0 is a constant
It can be shown that
PX ( x) = 1
x =0
and E(X) =
Poisson random variables are good for counting frequency of occurrence:
like the number of calls that arrive to a switchboard in one hour (busy
hour), or the number of packets that arrive to a router in one second.
246
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Gaussian/Normal distribution
Normal distribution:
Completely characterized by
mean () and variance (2)
247
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Gaussian distribution
248
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Error functions
Error functions
249
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Error functions
250
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Combining, we get:
Note: we can express Q and Phi functions with any real-valued arguments in terms
of the Q function with positive arguments alone.
Simplifies computation, enables use of bounds and approximations
for Q functions with positive arguments.
The rapid decay of the Q function implies, for example, that: Q(1) + Q(4) Q(1)
Design implications: We will use this to identify dominant events causing errors
251
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Gaussian distribution
252
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Gaussian distribution
Gaussian R.V.
Standard Gaussian :
Tail: Q(x)
tail decays exponentially!
253
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
X
Z=
Normal Standardized normal
distribution distribution
=1
X = 0 Z
One table!
Standardized normal
probability table (portion)
Z .00 .01 .02 =1
0.0 .0000 .0040 .0080
.0478
0.1 .0398 .0438 .0478
0.2 .0793 .0832 .0871
= 0 .12 Z
0.3 .1179 .1217 .1255 Shaded area
Probabilities exaggerated
254
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: P(X 8)
X 85
Z= = = .30
10
Normal Standardized Normal
Distribution Distribution
= 10 =1
.5000
.3821
.1179
=5 8 X =0 .30 Z
Shaded area exaggerated
Q-function:
Tail of normal
distribution
255
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Central tendency
Population distribution
x =
= 10
Dispersion
x =
n = 50 X
Sampling distribution
Sampling with
n=4 n =30
replacement X = 5 X = 1.8
X- = 50 X
As sample
size gets
large
enough
(n 30) ...
256
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
As sample x =
size gets n
sampling
large distribution
enough becomes
(n 30) ... almost normal.
X
x =
Comment on CLT
Central limit theorem works if original distribution are not
heavy tailed
Need to have enough samples.
E.g. with multipaths, if there is not rich enough
scattering, the convergence to normal may have not
happened yet.
Moments converge to limits.
Trouble with aggregates of heavy tailed distribution
samples
Rate of convergence to normal also varies with
distributional skew, and dependence in samples
257
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
258
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
259
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Covariance Matrix (m x m)
(i,j)th entry
Compact
representation
260
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Properties of covariance
Covariance unaffected when we add constants
261
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
262
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Often we deal with a linear combination (e.g., sample at output of a filter), which
are simply scalar Gaussian, so we do not need multidimensional integration.
263
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
264
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
265
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Related distributions
266
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Random processes
267
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
268
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Random processes
Random process
A random process is a collection of time functions, or signals,
corresponding to various outcomes of a random experiment.
For each outcome, there exists a deterministic function, which is called a
sample function or a realization.
Random
variables
Real number
Sample functions
or realizations
(deterministic
function)
time (t)
269
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
270
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
tn
t2
t0 t1
Random processes
271
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
272
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Stationarity
If time-shifts (any value T) do not affect its joint CDF
tn+T
tn
t2 t +T t2+T
t0 + T 1
t0 t1
273
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
With LTI systems, we will see that WSS inputs lead to WSS
outputs,
In particular, if a WSS process with PSD SX(f) is passed through a linear
time-invariant filter with frequency response H(f), then the filter output
is also a WSS process with power spectral density |H(f)|2SX(f).
Stationarity: Summary
Strictly stationary: If none of the statistics of the random
process are affected by a shift in the time origin.
274
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
275
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Correlation
276
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
277
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Ergodicity
Time averages = Ensemble averages
[i.e. ensemble averages like mean/autocorrelation can be computed as time-
averages over a single realization of the random process]
A random process: ergodic in mean and autocorrelation (like w.s.s.) if
and
278
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Power spectrum
For a deterministic signal x(t), the spectrum is well defined: If X ( )
represents its Fourier transform, i.e., if
+
X ( ) = x(t )e jt dt ,
then | X ( ) |2 represents its energy spectrum.
This follows from Parsevals theorem since the signal energy is given by
+ +
x (t )dt = 21 | X ( ) | d = E.
2 2
0 t 0
+
279
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Power signals:
Random process:
Power spectral density (PSD):
280
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Autocorrelation: Summary
Autocorrelation of an energy signal
Deterministic signals:
Random signals:
281
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Deterministic systems
Y ( t ) = g [ X ( t )]
Linear-Time Invariant
(LTI) systems
+
X (t ) h(t ) Y ( t ) = h ( t ) X ( ) d
+
LTI system = h ( ) X ( t ) d .
282
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
LTI systems
WSS input is good enough
Noise
283
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
[w/Hz]
Power spectral
density
Autocorrelation
function
Probability density function
284
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Approximation:
log2(1+x) x for small x
285
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Why?
Gaussian random process characterized by second order stats.
If second order stats are shift-invariant, then we cannot distinguish
statistically between shifted versions of the random process.
Sn ( f ) = N 0 /2 = 2 Rn ( ) = (N 0 /2) ( ) = 2 ( )
Two-sided PSD
(need to integrate over both positive and negative frequencies to get the power)
286
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
287
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Boltzmanns constant
288
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
289
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
White noise
Example
290
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
291
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
where
Need to learn some detection theory first, before we can deal with
this hypothesis testing problem.
Likelihood principle
Experiment:
Pick Urn A or Urn B at random
Select a ball from that Urn.
The ball is black.
What is the probability that the selected Urn is A?
292
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Likelihood principle
Likelihood principle
Bayes manipulations:
P(Urn A | Black) = P(Urn A and Black) /P(Black)
Decompose the numerator and denomenator in terms of the probabilities we know.
293
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
294
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Maximum Likelihood
Max likelihood
295
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Conditional densities
Prior probabilities
296
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Decision rule
A decision rule is a mapping from the observation space to the set of hypotheses
1
5
2 4 3
0 or 1 sent
Conditional densities
297
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
( )
298
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Check: Sensible rule for the basic Gaussian example is the ML rule
ML rule seems like a good idea. Is there anything optimal about it?
299
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
300
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Irrelevant statistics
301
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Irrelevant statistics
BEGIN PROOF
Conditional densities are all that are relevant. Under the given conditions,
These depend on hypothesis only through the first observation. END PROOF
Why?
302
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
303
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Decision threshold
304
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
[W/Hz]
Power spectral
Density
(flat => white)
Autocorrelation
Function
Probability density function (uncorrelated)
(gaussian)
305
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
n
mi si z m
Modulator Decision rule
pz (z | m2 ) pz (z | m1 )
s2 s1
1 (t )
Eb 0 Eb
306
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
likelihoods
307
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
pz (z | m2 ) pz (z | m1 )
s2 s1
1 (t )
Eb 0 Eb
s s /2
Pe ( m1 ) = Pe (m2 ) = Q 1 2
N /2
0
2 Eb
PB = PE ( 2) = Q
N0
308
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Vector detection
309
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
310
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
LMMSE
622
311
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
623
Vector
312
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
625
Line equation: y = mx + c
Matrix equation: y = Mx + c
626
313
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
A
A+B = C
(use the head-to-tail method
B to combine vectors)
C
B
627
Scalar product: av
av = a( x1 , x2 ) = (ax1 , ax2 )
av
v
314
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
phase
x
629
v
w v.w = ( x1 , x2 ).( y1 , y2 ) = x1 y1 + x2 . y2
v.w = 0 v w
If vectors v, w are columns, then dot product is wTv
630
315
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
= cross-correlation between x(t) and y(t)
Properties of inner product:
< 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 ) >
631
Signal space
The distance in signal space is measure by calculating the norm.
What is norm?
Norm of a signal (generalization of length):
x(t ) = < x(t ), x(t ) > = x(t ) dt = E x
2
= length of x(t)
ax(t ) = a x(t )
d x , y = x(t ) y (t )
316
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
E1 d s1 , z
1 (t )
E3 z = ( z1 , z 2 )
d s3 , z E2 d s2 , z
s 3 = (a31 , a32 )
s 2 = (a21 , a22 )
Detection in
The Euclidean distance between signals z(t) and s(t): AWGN noise:
d si , z = si (t ) z (t ) = (ai1 z1 ) 2 + ( ai 2 z2 ) 2 Pick the closest
i = 1,2,3 signal vector
633
y = [0 1 0] x z = 0
Note: T
317
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
CDMA codes are orthogonal, and projecting the composite received signal
on each code helps extract the symbol transmitted on that code. 635
Code B
Code A
B
B
Code A A
A
B C C
B B C
A A A B
A C
B
Time
Sender Receiver
636
Each code is an orthogonal basis vector signals sent are orthogonal
318
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Matrix
a b
c d
columns
637
Matrix (geometrically)
Matrix represents a linear function acting on vectors:
Linearity (a.k.a. superposition): f(au + bv) = af(u) + bf(v)
f transforms the unit x-axis basis vector i = [1 0]T to [a c]T
f transforms the unit y-axis basis vector j = [0 1]T to [b d]T
f can be represented by the matrix with [a c]T and [b d]T as columns
Why? f(w = mi + nj) = A[m n]T
Column viewpoint: focus on the columns of the matrix!
a b
c d
[0,1]T [a,c]T
[1,0]T
[b,d]T
Linear Functions f : Rotate and/or stretch/shrink the basis vectors
319
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
a b x x'
x = ax + by
=
c d y y'
y = cx + dy
Note: Linear channel codes (eg: Hamming, Reed-Solomon, BCH) can be viewed as
k-dimensional vector sub-spaces of a larger N-dimensional space.
k-data bits can therefore be protected with N-k parity bits
640
320
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
FEC (N-K)
Block
Size Lossy Network
(N)
Data = K
r1 0
0 r2
[0,1]T [0,r2]T
scaling
[1,0]T [r1,0]T
cos -sin
sin cos
[-sin, cos]T
[0,1]T
[cos, sin]T
rotation
[1,0]T 642
321
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Scaling
P
Rotation
cos -sin
sin cos
644
322
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Reflections
Reflection can be about any line or point.
Complex Conjugate: reflection about x-axis
(i.e. flip the phase to -)
Reflection => two times the projection
distance from the line.
Reflection does not affect magnitude
Induced Matrix
645
646
323
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
2D Translation
P
t
P
P
P ' = ( x + t x , y + t y ) = P+t ty
t
P
y
x tx 647
a b e f a + e b + f
c d + g =
h c + g d + h
Just add elements
a b e f a e b f
c d g =
h c g d h
Just subtract elements
a b e f ae + bg af + bh
c d g = Multiply each row by
h ce + dg cf + dh each column
648
324
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Multiplication
Is AB = BA? Maybe, but maybe not!
a b e f ae + bg ... e f a b ea + fc ...
c d g = =
h ... ... g
h c d ... ...
649
Multiplication as composition
Different!
650
325
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Inverse of a matrix
Identity matrix:
AI = A
Inverse exists only for square
matrices that are non-singular
Maps N-dim space to
1 0 0
I = 0 1 0
another N-dim space
bijectively
Some matrices have an
inverse, such that:
AA-1 = I
0 0 1
Inversion is tricky:
(ABC)-1 = C-1B-1A-1
Determinant of a matrix
a b
Used for inversion A=
If det(A) = 0, then A has no inverse c d
det( A) = ad bc
Note: Determinant-criterion for space-time
code design.
Good code exploiting time diversity 1 d b
A1 =
ad bc c a
should maximize the minimum
product distance between codewords.
Coding gain determined by min of
determinant over code words.
652
326
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
p = a (aTx)
||a|| = aTa = 1
653
Note: We can use this idea to find a least-squares line that minimizes
the sum of squared errors (i.e. min eTe).
This is also used in detection under AWGN noise to get the test statistic:
Idea: project the noisy received vector y onto (complex) transmit vector h:
matched filter/max-ratio-combining (MRC)
654
327
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
655
Signal + AWGN noise will not reveal the original transmitted sequence.
There is a high power of noise relative to the power of the desired signal (low SNR).
If the receiver were to sample this signal at the correct times, the
resulting binary message would have a lot of bit errors.
656
328
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Matched filter
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
657
||h||
658
Multiply by conjugate => cancel phase!
329
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Symmetric: A = AT
Symmetric => square matrix
Complex vectors/matrices:
Transpose of a vector or a matrix with complex elements must involve a
conjugate transpose, i.e. flip the phase as well.
For example: ||x||2 = xHx, where xH refers to the conjugate transpose of x
Hermitian (for complex elements): A = AH
Like symmetric matrix, but must also do a conjugation of each element (i.e. flip
its phase).
i.e. symmetric, except for flipped phase
Note we will use A* instead of AH for convenience
Positive definite: symmetric, and its quadratic forms are strictly positive, for non-
zero x :
xTAx > 0
Geometry: bowl-shaped minima at x = 0
659
Note:
Gaussian noise exhibits isotropy, i.e. invariance to direction. So any rotation
Q of a gaussian vector (w) yields another gaussian vector Qw.
Circular symmetric (c-s) complex gaussian vector w => complex rotation with
U yields another c-s gaussian vector Uw
Note: The Discrete Fourier Transform (DFT) matrix is both unitary and symmetric.
DFT is nothing but a complex rotation, i.e. viewed in a basis that is a rotated
version of the original basis.
FFT is just a fast implementation of DFT. It is fundamental in OFDM.
660
330
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note: Gaussian vector formula has a quadratic form term in its exponent:
exp[-0.5 (x -)T K-1 (x -)]
Similar to 1-variable gaussian: exp(-0.5 (x -)2/2 )
K-1 (inverse covariance matrix) instead of 1/ 2
Quadratic form involving (x -) instead of (x -)2
661
Rectangular matrices
Linear system of equations:
Ax = b
More or less equations than necessary.
Not full rank
If full column rank, we can modify equation as:
ATAx = ATb
Now (ATA) is square, symmetric and invertible.
x = (ATA)-1 ATb now solves the system of equations!
This solution is called the least-squares solution. Project b onto column space
and then solve.
(ATA)-1 AT is sometimes called the pseudo inverse
Note: (ATA) or (A*A) will appear often in communications math (MIMO). They
will also appear in SVD (singular value decomposition)
The pseudo inverse (ATA)-1 AT will appear in decorrelator receivers for MIMO
662
331
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
E-vectors:
- Points on the x-axis unaffected [1 0]T
- Points on y-axis are flipped [0 1]T
(but this is equivalent to scaling by -1!)
E-values: 1, -1 (also on diagonal of matrix)
663
Eigenvectors
Eigenvectors are even more interesting because any vector in the domain of
T can now be viewed in a new coordinate system formed with the invariant
eigen directions as a basis.
The operation of T(x) is now decomposable into simpler operations on x,
which involve projecting x onto the eigen directions and applying the
characteristic (eigenvalue) scaling along those directions
664
332
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
The eigenvectors
1 and 1 form U = 1 1
1 1 1
1
1 1 / 2 1 / 2 Recall
Inverting, we have U = UU1 =1.
1 / 2 1 / 2
1 1 1 0 1 / 2 1 / 2
Then, S=UU1 = 1 1 0 3 1 / 2 1 / 2
666
333
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
1 / 2 1 / 2 1 0 1 / 2 1/ 2
Then, S=
1 / 2 1 / 2 0 3 1 / 2 1/ 2
Q (Q-1= QT )
667
where
Eigenvector decomposition:
334
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
669
Eigen decomposition
Every square matrix A, with distinct eigenvalues has an eigen
decomposition:
A = SS-1
S is a matrix of eigenvectors and
is a diagonal matrix of distinct eigenvalues = diag(1, N)
670
335
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
This is why we like symmetric (or hermitian) matrices: they admit nice
decomposition
We like positive definite matrices even more: they are symmetric and
all have all eigenvalues strictly positive.
Many linear systems are equivalent to symmetric/hermitian or positive
definite transformations.
671
336
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Fourier/Eigen decomposition
Continuous case:
Any function f(t) can be viewed as a integral (sum) of
scaled, time-shifted impulses c()(t+) d
h(t) is the response the system gives to an impulse (impulse
response).
Functions response is the convolution of the function f(t)
with impulse response h(t): for linear time-invariant systems
(LTI): f(t)*h(t)
Convolution is messy in the time-domain, but becomes a
simple multiplication in the frequency domain: F(s)H(s)
Input Output
Linear system
673
337
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
338
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Digital Modulation
Basic concepts
Modulation
Placing baseband signals on high frequency carriers using the
process of modulation facilitates the long distance
transmission of data, voice and video signals.
Modulation:
The signal processing technique where, at the transmitter
one signal (the modulating signal) modifies a property of
another signal (the carrier signal) so that a composite wave
(the modulated wave) is formed.
Demodulation:
At the receiver, the modulating signal is recovered from the
modulated wave.
678
339
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Bandwidth
679
Analogue modulation
680
340
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
681
682
341
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Industry trends:
683
684
342
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Transmitting information
Modulation
685
Polar display
686
343
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Polar display
687
I/Q formats
In digital communications,
modulation is often expressed in
terms of I and Q.
This is a rectangular representation
of the polar diagram.
On a polar diagram, the I axis lies
on the zero degree phase reference,
and the Q axis is rotated by 90
degrees.
The signal vectors projection onto
the I axis is its I component and
the projection onto the Q axis is its
Q component.
688
344
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
I and Q in transmitter
689
Transmitter side
345
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Receiver side
691
692
346
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Application areas
Modulation format Application
MSK, GMSK GSM
BPSK Deep space telemetry, cable modems
QPSK and DQPSK Satellite, CDMA, TETRA
OQPSK (OffsetQPSK) CDMA, satellite
FSK, GFSK DECT, paging, AMPS, CT2,
Terrestrial Trunked Radio (TETRA) is a professional mobile radio and two-way transceiver
specification. TETRA was specifically designed for use by government agencies, emergency
services, (police forces, fire departments, ambulance), transport services and the military.
693
TETRA is an European Telecommunications Standards Institute (ETSI) standard.
Digital modulation
694
347
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
695
696
348
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
697
698
349
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Bipolar operation:
A 1 is represented by a current of +A signal units
A 0 is represented by a current of A signal units
Two modes of operation, once again:
Non Return to Zero
Currents maintained for entire time slot
Power needed for equally likely symbols is A2 signal watts
Return to Zero
Currents maintained for fraction of time slot
Power needed for equally likely symbols is A2/2 signal watts
699
700
350
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
701
702
351
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
703
101100111000
704
352
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
2B1Q signaling is used for BISDN basic rate services (at 160kbps) and
ISDN
digital subscriber loop services.
For long sequences of 1s and 0s, or alternating 1s and 0s (i.e. 1010101010)
2B1Q
signaling produces constant currents and synchronization is impossible.
Since the frequency power density spectrums of 2B1Q, AMI and Raised
Cosine
are narrower, they are employed in bandwidth limited environments such as
telephone connections.
Manchester is used in LANs and other applications where precise
synchronization is important and bandwidth is available.
705
Bit Rate
706
353
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
707
Bandwidth requirements
BPSK 8PSK
One bit per symbol 3 bits per symbol
Bit rate = Symbol rate Symbol rate = 1/3 Bit rate
708
354
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
709
Symbol clock
The symbol clock represents the frequency and exact
timing of the transmission of the individual symbols.
At the symbol clock transitions, the transmitted carrier
is at the correct I/Q (or magnitude/phase) value to
represent a specific symbol (a specific point in the
constellation).
710
355
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
711
Binary keying
712
356
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1 is present as:
713
714
357
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
OOK
On-off keying is also known as Amplitude Shift Keying (ASK)
The above graph shows a time domain representation of Binary
Amplitude Shift Keying
715
716
358
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
This is illustrated in the chart above. Notice the 180 phase shifts
indicated by the arrow.
717
718
359
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
g (t ) = jAc m(t )
719
720
360
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
721
722
361
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
723
724
362
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
725
726
363
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
/4-QPSK
This final variant of QPSK uses two
identical constellations which are
rotated by 45 ( / 4 radians, hence
the name) with respect to one another.
Usually, either the even or odd symbols
are used to select points from one of
the constellations and the other
symbols select points from the other
constellation.
This also reduces the phase-shifts
from a maximum of 180, but only to a
maximum of 135 and so the amplitude
fluctuations of / 4QPSK are
between OQPSK and non-offset QPSK.
One property this modulation scheme
possesses is that if the modulated
signal is represented in the complex
domain, it does not have any paths
through the origin.
In other words, the signal does not
pass through the origin.
Dual constellation diagram for /4-QPSK.
This lowers the dynamical range of
fluctuations in the signal which is This shows the two separate constellations
desirable when engineering with identical Gray coding but rotated by
communications signals. 45 with respect to each other
QPSK Summary
Quadrature means that the signal shifts between
phase states that are separated by 90 degrees (/2
radians). The signal shifts in increments of 90 degrees
from 45 to 135, 45, or 135 degrees.
These points are chosen as they can be easily
implemented using an I/Q modulator.
Only two I values and two Q values are needed and
this gives two bits per symbol.
There are four states because 22 = 4.
It is therefore a more bandwidth-efficient type of
modulation than BPSK - twice as efficient.
728
364
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
729
730
365
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
731
FSK
where
732
366
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
733
734
367
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
735
736
368
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
737
738
369
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
739
740
370
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Differential PSK
741
DPSK
Differential Phase-Shift Keying
Binary data are first differentially encoded and then passed to
the BPSK modulator.
Example:
742
371
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
DPSK
743
Example: DSPK
744
372
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Digital Modulation
Information:
Information
- analog:BW &
source
dynamic range Message Message
- digital:bit rate estimate Information
sink
Source
Maximization of encoder Source
decoder
information
transferred Channel Channel In baseband systems
Encoder decoder these blocks are missing
Message protection &
channel adaptation; Baseband
Interleaving Deinterleaving means that
convolution, block
coding no carrier
Modulator Demodulator wave
Fights against burst modulation
errors is used for
Transmitted Received signal transmission
Channel (may contain errors)
M-PSK/FSK/ASK..., signal
depends on channel wired/wireless
Noise constant/variable
BW & characteristics 746
Interference linear/nonlinear
373
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
747
Modulation
Transform digital data into an analog signal that
can be transmitted or stored (the real world is
analog, not digital).
Demodulation/detection
The received signal contains information about the
transmitted data but is corrupted by noise.
Estimate what data was sent, aiming at minimum
possible probability of making mistakes.
748
374
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
749
Modulation
750
375
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Modulation
751
752
376
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
753
Basis waveforms
754
377
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
755
Signal space
756
378
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
757
Linear independency
758
379
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
759
Gram-Schmidt
760
380
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
761
Example: PAM
762
381
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
763
Example: PPM
764
382
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: PPM
765
766
383
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: PSK
767
Example: PSK
768
384
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
769
770
385
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
771
Modulation/demodulation
772
386
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Receiver
773
774
387
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Optimal demodulation
775
Correlator demodulator
776
388
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
777
Matched filtering+sampling
Correlation
778
389
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
779
780
390
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
781
Example
782
391
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Error probability
783
784
392
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
785
Q-function
786
393
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Examples
787
Example: OOK
788
394
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
789
790
395
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Pe vs. SNR
791
792
396
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
793
794
397
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
795
Note
Union Bound:
A B
Applications:
Getting bounds on BER ,
In general, bounding the tails of probability distributions 796
398
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
797
Digital Modulation
798
399
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
799
800
400
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
801
802
401
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
803
G-S orthogonalization
804
402
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
G-S orthogonalization
805
Example
806
403
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example (contd.)
807
Example (contd.)
808
404
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: Summary
809
810
405
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
811
Note
812
406
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
813
814
407
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
815
M-ary PSK
816
408
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
M-ary PSK
817
M-ary PSK
818
409
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
M-ary PSK
819
M-ary QAM
820
410
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
M-ary QAM
821
M-ary QAM
Signal space diagram (M=16)
822
411
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
M-ary QAM
Signal space diagram (M=8)
823
M-ary QAM
824
412
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
M-ary FSK
825
M-ary FSK
826
413
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
M-ary FSK
Signal space diagram (M = 2)
827
M-ary FSK
828
414
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
830
415
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Advantages of OFDM
The primary advantage of OFDM over single-carrier schemes is its ability
to cope with severe channel conditions (for example, attenuation of high
frequencies in a long copper wire, narrowband interference and
frequency-selective fading due to multipath) without complex
equalization filters.
Channel equalization is simplified because OFDM may be viewed as
using many slowly-modulated narrowband signals rather than one rapidly-
modulated wideband signal.
The low symbol rate makes the use of a guard interval between symbols
affordable, making it possible to handle time-spreading and eliminate
intersymbol interference (ISI).
This mechanism also facilitates the design of single frequency networks
(SFNs), where several adjacent transmitters send the same signal
simultaneously at the same frequency, as the signals from multiple distant
transmitters may be combined constructively, rather than interfering as
would typically occur in a traditional single-carrier system.
831
Multicarrier modulation
As it is known from the single carrier based communication
systems nonideal channels introduce Intersymbol Interference
(ISI), which degrades the performance compared with the ideal
channel.
The degree of performance degradation depends on the
frequency response of the channel.
Typically, the complexity of the receiver increases as the
spread of the ISI increases.
An alternative approach to the design of bandwidth-efficient
communication system in the presence of channel distortion is
to subdivide the available channel bandwidth into a number of
narrow sub-channels so, that the frequency response of each
subchannel is nearly flat.
Multicarrier modulation method has been used in variety of
applications (e.g., DAB, DVB-T/H, 3GPP-LTE, ADSL and
HDSL). 832
416
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
833
834
417
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
835
P distribution
836
418
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
837
Transmitter
Idealized system model:
A simple idealized OFDM system model suitable for a time-invariant AWGN
channel transmitter
838
419
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Transmitter
An OFDM carrier signal is the sum of a number of orthogonal sub-carriers, with
baseband data on each sub-carrier being independently modulated commonly using
some type of quadrature amplitude modulation (QAM) or phase-shift keying
(PSK).
This composite baseband signal is typically used to modulate a main RF carrier.
Input signal s[n] is a serial stream of binary digits.
By inverse multiplexing, these are first demultiplexed into N parallel streams, and
each one mapped to a (possibly complex) symbol stream using some modulation
constellation (QAM, PSK, etc.).
Note that the constellations may be different, so some streams may carry a higher
bit-rate than others.
An inverse FFT (IFFT) is computed on each set of symbols, giving a set of complex
time-domain samples.
These samples are then quadrature-mixed to passband in the standard way.
However, the real and imaginary components are first converted to the analogue
domain using digital-to-analogue converters (DACs); the analogue signals are then
used to modulate cosine and sine waves at the carrier frequency, fc.
These signals are then summed to give the transmission signal s(t).
839
Receiver
The receiver picks up the signal r(t) , which is then
quadrature-mixed down to baseband using cosine and sine
waves at the carrier frequency.
This also creates signals (mirror-images) centered on 2fc , so
low-pass filters are used to reject these.
The baseband signals are then sampled and digitised using
analogue-to-digital converters (ADCs).
Next FFT is used to convert signals back to the frequency
domain.
This returns N parallel streams, each of which is converted to a
binary stream using an appropriate symbol detector.
These streams are then re-combined into a serial stream, which
is an estimate of the original binary stream at the transmitter.
840
420
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Receiver
Idealized system model:
A simple idealized OFDM system model suitable for a time-invariant
AWGN channel receiver
841
842
421
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Multicarrier spectrum
843
Spectra
With rectangular pulse shaping, the amplitude spectrum of the subcarrier is
844
422
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
OFDM transmitter
845
846
423
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
OFDM Receiver
The multiplexing operations in receiver can be implemented by using DFT
(Discrete Fourier Transform) operations
In practice, DFT is implemented by using FFTs (Fast Fourier Transforms).
847
848
424
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
849
850
425
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Orthogonality
In OFDM, the sub-carrier frequencies are chosen so that the sub-carriers
are orthogonal to each other, meaning that cross-talk between the sub-
channels is eliminated and inter-carrier guard bands are not required.
This greatly simplifies the design of both the transmitter and the receiver;
unlike conventional FDM, a separate filter for each sub-channel is not
required.
The orthogonality requires that the sub-carrier spacing is f=k/TU Hz,
where TU seconds is the useful symbol duration (the receiver side window
size), and k is a positive integer, typically equal to 1.
Therefore, with N sub-carriers, the total passband bandwidth will be B
Nf (Hz).
The orthogonality also allows high spectral efficiency, with a total symbol
rate near the Nyquist rate for the equivalent baseband signal (i.e. near half
the Nyquist rate for the double-side band physical passband signal),
because almost the whole available frequency band can be utilized.
OFDM generally has a nearly 'white' spectrum, giving it gentle
electromagnetic interference properties with respect to other co-channel
users.
851
852
426
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
853
Avoiding ISI
854
427
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Channel responses
856
428
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Channel estimation
857
858
429
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
If one sends a million symbols per second using conventional single-carrier
modulation over a wireless channel, then the duration of each symbol
would be one microsecond or less.
This imposes severe constraints on synchronization and necessitates the
removal of multipath interference.
If the same million symbols per second are spread among one thousand
sub-channels, the duration of each symbol can be longer by a factor of a
thousand (i.e., one millisecond) for orthogonality with approximately the
same bandwidth.
Assume that a guard interval of 1/8 of the symbol length is inserted
between each symbol.
Intersymbol interference can be avoided if the multipath time-spreading
(the time between the reception of the first and the last echo) is shorter than
the guard interval (i.e., 125 microseconds).
This corresponds to a maximum difference of 37.5 kilometers between the
lengths of the paths.
859
860
430
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
861
Simplified equalization
The effects of frequency-selective channel conditions, for
example fading caused by multipath propagation, can be
considered as constant (flat) over an OFDM sub-channel if the
sub-channel is sufficiently narrow-banded (i.e., if the number
of sub-channels is sufficiently large).
This makes equalization far simpler at the receiver in OFDM
in comparison to conventional single-carrier modulation.
The equalizer only has to multiply each detected sub-carrier
(each Fourier coefficient) by a constant complex number, or a
rarely changed value.
862
431
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
A low-density parity-check (LDPC) and turbo codes are capacity-approaching codes, which means that
864
practical constructions exist that allow codes to closely approach the channel capacity, a theoretical
maximum for the code rate at which reliable communication is still possible given a specific noise level.
432
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
865
Example
866
433
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Introduction to Information
and Coding Theory
Basic concepts
868
434
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Information:
Information
- analog:BW &
source
dynamic range Message Message
- digital:bit rate estimate Information
sink
Source
Maximization of encoder Source
information decoder
transferred Channel Channel In baseband systems
Encoder decoder these blocks are missing
Message protection &
channel adaptation; Baseband
Interleaving Deinterleaving means that
convolution, block
coding no carrier
Modulator Demodulator wave
Fights against burst modulation
errors is used for
Received signal transmission
Transmitted
signal Channel (may contain errors)
M-PSK/FSK/ASK...,
depends on channel wired/wireless
Noise constant/variable
BW & characteristics 869
Interference linear/nonlinear
870
435
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
P(not a) = 1 P(a)
P(a or b) = P(a)+P(b) P(a and b),
where a and b are events
We will often denote P(a and b) by P(a, b).
If P(a,b) = 0, we say a and b are mutually exclusive.
871
Conditional probability
P(a|b)P(b) = P(b|a)P(a),
or
P(a|b) =P(b|a)P(a)/P(b)
872
436
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Indepencence
If two events a and b are such that
P(a|b) = P(a),
we say that the events a and b are independent.
Note that from Bayes Theorem, we will also have
that
P(b|a) = P(b),
and furthermore,
P(a,b) = P(a|b)P(b) = P(a)P(b).
This last equation is often taken as the definition of
independence.
873
874
437
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
4. And thus, by continuity, we get, for 0 < p 1, and a real number a > 0 :
I(pa) = a I(p)
875
Information measure
We can find a simple expression, which satisfy the previous
properties.
This is
I(p) = logb(p) = logb(1/p)
for any base b.
The base b determines the units we are using.
We can change the units by changing the base, using the
formulas, for bases b1, b2, x > 0,
and therefore
876
438
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Units of information
Thus, using different bases for the logarithm results in
information measures which are just constant multiples of each
other, corresponding with measurements in different units:
log2 units are bits (from binary)
loge units are nats from natural logarithm
log10 units are hartleys, after an early scientist in the field of
transmission techniques.
877
Example
878
439
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
879
880
440
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
881
Note
Note that
882
441
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
883
884
442
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Definition of information
Here the message signal is modeled as a random process.
We begin by considering observations of a random variable:
Each observation gives a certain amount of information.
But rare observations give more information than usual ones.
885
Definition: (Self)information
I(xm) = -log2(pX(xm))=log2(1/pX(xm))
886
443
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Interpretation
= I(xm)+ I(yn)
Thus in case of independent events, the information is additive, which makes sense
intuitively.
887
Information source
888
444
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
889
890
445
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Entropy
( )
891
Entropy
Entropy has the following interpretations:
Average information obtained from an observation.
Average uncertainty about X before the observation.
Entropy is measure of uncertainty.
The more we know about something the lower the entropy.
892
446
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
893
894
447
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
895
Example
896
448
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Theorem
We have H(X) = 0 when exactly one of the probabilities is one and all
the rest are zero. We have H(X) = log(L) only when all of the events
have the same probability 1/L. That is, the maximum of the entropy
function is the log() of the number of possible events, and occurs
when all the events are equally likely.
897
Example
How much information can a student get from a
single grade?
First, the maximum information occurs if all grades
have equal probability.
E.g., in a pass/fail class, on average half should pass
if we want to maximize the information given by the
grade.
The maximum information the student gets from a
grade will be:
Pass/Fail : 1 bit.
Pass: 1, 2, 3, 4, 5 : 2.3 bits.
898
449
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
899
Average entropy
450
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note
Comment
It is important to recognize that our definitions of information
and entropy depend only on the probability distribution.
In general, it would not make sense for us to talk about the
information or the entropy of a source without specifying the
probability distribution.
It can certainly happen that two different observers of the same
data stream have different models of the source, and thus
associate different probability distributions to the source.
The two observers will then assign different values to the
information and entropy associated with the source.
902
451
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example on comment
Conditional entropy
Now given two random variables X and Y, the
conditional entropy of X given Y is denoted as
H(X|Y) and measures
average information obtained from observing X given that
the value of Y is known
average uncertainty about the observation X given that the
value of Y is known
how much extra information one still needs to supply on
average to communicate X given that the other party knows
Y
Thus the conditional entropy measures the statistical
dependence between X and Y in information theoretic
sense.
904
452
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Conditional entropy
905
cf. confer
Conditional entropy
906
453
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Theorem
907
454
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Chain rule
909
Example
910
455
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
911
Example
912
456
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: Dice
913
Example: Dice
914
457
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Theorem
915
Entropy rate
916
458
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
917
918
459
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
919
Mutual information
460
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Note
=> I(X,Y) = 0
921
Example
922
461
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
923
924
462
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Information measures
925
926
463
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
927
928
464
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
929
Transmission channels
Cables
wire pairs (e.g., ordinary telephone line)
coaxial cable
waveguide (metallic waveguide and optical fiber)
More or less free space radio transmission
broadcasting
point-to-point microwave transmission
satellite position transmission
cell networks
(Portable magnetic/electronic/optical memory equipment)
930
465
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Channel models
Examples:
A discrete channel A linear additive noise channel
932
466
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
933
Example
Well assume, for simplicity, that the channel is time-invariant and
memoryless, so the conditional probabilities are independent of time and
previous symbol transmissions.
The conditional probabilities P(yj|xi) then have special significance as the
channels forward transition probabilities.
By way of example, Figure depicts the forward transitions for a noisy
channel with two source symbols and three destination symbols.
If this system is intended to deliver yj = y1 when xi = x1 and yj = y2 when xi
= x2 , then the symbol error probabilities are given by P(yj|xi) for j i.
Forward
transition
probabilities for
a noisy discete
channel
934
467
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Shannons model
In Shannons discrete model, it is assumed that the
source provides a stream of symbols selected from a
finite alphabet A = {a1, a2, . . . , an}, which are then
encoded.
The code is sent through the channel and possibly
disturbed by noise.
At the other end of the channel, the receiver will
decode, and derive information from the sequence of
symbols.
Shannons model
Given a source of symbols and a channel with noise (more precise, a
probability model for these elements), we can talk about the capacity of the
channel.
The general model Shannon worked with involved two sets of symbols, the
input symbols and the output symbols.
Let us say the two sets of symbols are
A = {a1, a2, . . . , an} and
B = {b1, b2, . . . , bm}.
Note that we do not necessarily assume the same number of symbols in the
two sets.
Given the noise in the channel, when symbol bj comes out of the channel,
we can not be sure which ai was put in.
The channel is characterized by the set of probabilities {P(ai|bj)}.
936
468
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Mutual information
We can then consider various related information and entropy
measures.
First, we can consider the information we get from observing a
symbol bj.
Given a probability model of the source, we have an a priori
estimate P(ai) that symbol ai will be sent next.
Upon observing bj, we can revise our estimate to P(ai|bj).
The change in our information (the mutual information) will
be given by:
937
938
469
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
939
Also we have:
I(A;B) 0,
and
I(A;B) = 0
if and only if A and B are independent.
940
470
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
941
942
471
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Channel capacity
943
944
472
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
945
946
473
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
(the joint
947
948
474
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Possible sequences
949
Coding problem
475
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
476
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
953
Typical sequence
954
477
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
956
478
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
957
Typical set
958
479
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
959
960
480
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
Choose a smaller , namely =0.046 [ 5% of
h(1/3)], and increase the length of the
sequences.
961
Meaningful sequences
481
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Meaningful text
963
Meaningful fraction
964
482
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
965
966
483
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
967
Codewords
968
484
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Code rate
969
Transmission rate
485
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
971
972
486
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
973
974
487
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
975
976
488
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
By combining
and the previous equations we obtain
977
978
489
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Fan
979
Fans
980
490
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Fans
981
982
491
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Channel capacity
983
984
492
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
R>C
985
986
493
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Reversed fans
987
Coding strategy
988
494
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
989
990
495
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
991
992
496
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
993
994
497
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
995
996
498
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
997
998
499
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
999
1000
500
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Channel capacity
1001
Channel capacity
1002
501
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1003
1004
502
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1005
Energy/bit
1006
503
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1007
Shannon limit
1008
504
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1009
1010
505
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1011
1012
506
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1013
1014
507
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1015
1016
508
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Source coding
1017
1018
509
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1019
1020
510
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Classifications
1021
511
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Inequalites
1023
1024
512
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Source coding
1025
Source coding
1026
513
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Prefix
1027
Prefix-free code
1028
514
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
1029
Example
1030
515
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1031
1032
516
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1033
Example
1034
517
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1035
Example: Huffman
1036
518
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: Huffman
1037
Example: Huffman
1038
519
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1039
1040
520
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1041
Huffman codes
1042
521
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Huffman codes
1043
Lempel-Ziv-Welch
522
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Lempel-Ziv algoritm
1045
1046
523
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1047
Distortion measures
1048
524
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Rate-distortion function
1049
Distortion-rate function
1050
525
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Examples
1051
Quantization
1052
526
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Scalar quantizer
1053
1054
527
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Uniform/Linear quantization
1055
Quantization noise
1056
528
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1057
1058
529
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Speech compression
1059
1060
530
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1061
1-point DPCM
1062
531
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Multipoint DPCM
1063
1064
532
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Analysis-syntesis technique
1065
Some examples
1066
533
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1067
1068
534
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Binary field
1069
Addition
1070
535
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Linear code
1071
1072
536
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1073
Minimum distance
1074
537
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Table 1:
1075
Example
1076
538
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
1077
Example
1078
539
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
1079
Example
1080
540
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
1081
Example
1082
541
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
G-matrix
1083
Generation of codewords
1084
542
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
1085
Parity-checking procedure
1086
543
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Parity-checking procedure
1087
Parity-check matrix
1088
544
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
GHT=0
1089
Convolutional codes
1090
545
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1091
Encoder state
546
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example
1093
1094
547
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Trellis-diagram
1095
Viterbi-algorithm
1096
548
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1097
1098
549
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1099
1100
550
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1101
Diversity
551
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Introduction to diversity
Basic Idea
Send same bits over independent fading paths
Independent fading paths obtained by time, space, frequency, or polarization
diversity
Combine paths to mitigate fading effects
Tb
t
Multiple paths unlikely fade simultaneously
1103
Diversity gain
(MIMO):
552
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Time diversity
1105
Time diversity
Time diversity can be obtained by interleaving and coding over
symbols across different coherent time periods.
Channel: time
diversity/selectivity,
but correlated across
successive symbols
(Repetition) Coding
without interleaving: a full
codeword lost during fade
Interleaving: of sufficient
depth: (> coherence time)
At most 1 symbol of codeword
lost!
Coding alone is not sufficient!
553
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
FEC (N-K)
Block
Size Lossy Network
(N)
Data = K
1108
554
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Example: GSM
The data of each user are sent over time slots of length 577 s
Time slots of the 8 users together form a frame of length 4.615 ms
Voice: 20 ms frames, rate convolution coded = 456 bits/voice-frame
Interleaved across 8 consecutive time slots assigned to that specific user:
0th, 8th, . . ., 448th bits are put into the first time slot,
1st, 9th, . . ., 449th bits are put into the second time slot, etc.
One time slot every 4.615 ms per user, or a delay of ~ 40 ms (ok for voice).
The 8 time slots are shared between two 20 ms speech frames.
1109
where and
555
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
||h||
1111
Multiply by conjugate => cancel phase!
1112
556
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Chi-square distribution
In probability theory and statistics, the chi-square distribution
(also chi-squared or -distribution) with k degrees of freedom
is the distribution of a sum of the squares of k independent
standard normal random variables.
It is one of the most widely used probability distributions in
hypothesis testing, or in construction of confidence intervals
If X1, ..., Xk are independent, standard normal random variables,
then the sum of their squares
The chi-square distribution has one parameter: k a positive integer that specifies the
1113
number of degrees of freedom (i.e. the number of Xis)
Chi-Squared pdf of
1114
557
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Antenna diversity
1116
558
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Antenna diversity
1117
Antenna diversity: Rx
1118
559
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Receive diversity
1119
1120
560
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1121
SNR:
561
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
SNR: 1123
1124
562
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Antenna diversity: Tx
1125
Transmit diversity
1126
563
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1127
1128
564
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1129
Wireless Overview
565
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Transmitter
Source Channel
Formatter Modulator
encoder encoder
Receiver
Source Channel
Formatter Demodulator
decoder decoder
566
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Why wireless?
Characteristics
Mostly radio transmission
New protocols for data transmission are needed
Advantages
Spatial flexibility in radio reception range
Ad hoc networks without former planning
No problems with wiring (e.g. historical buildings, fire protection, esthetics)
Robust against disasters like earthquake, fire and careless users (which
remove and break connectors and cut wires)
Disadvantages
Generally lower transmission rates for higher numbers of users
Often proprietary, standards are often restricted
Many national regulations, global regulations are evolving slowly
Restricted frequency range, interferences of frequencies
Nevertheless, in the last 30 years, it has really been a wireless revolution 1133
567
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1135
568
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Fading channel
569
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Wireless systems
Cellular
With a big emphasis on voice communication
Terrestrial microwave and satellite systems
WiFi
Local networks over wireless, with infrastructure
E.g., 802.11a,b,g,n
WiMAX
Internet provider last mile replacement
Ad Hoc Network
Local networks over wireless, without infrastructure
Sensor network
Radar and radio telescope system 1139
Cellular systems
Geographic region divided into cells
Frequencies/timeslots/codes reused at spatially-separated
locations.
Base stations/MTSOs (Mobile Telephone Switching Offices)
coordinate handoff and control functions
Shrinking cell size increases capacity, as well as networking
burden
Note: Co-channel interference (between same-color cells
below).
BASE
STATION
MTSO
1140
570
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
BS
BS
Internet
New York
MTSO MTSO
PSTN
BS
1141
PSTN - Public ServiceTelephone Network
1142
571
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Wireless generations
First Generation (1G): Analog 25 or 30 kHz FM,
voice only, mostly vehicular communication
Second Generation (2G): Narrowband TDMA and
CDMA, voice and low bit-rate data, portable
units.
Third Generation (3G): Wideband TDMA and
CDMA, voice and high bit-rate data, portable
units
Fourth Generation (4G and beyond 2015): true
broadband wireless: advanced versions of
WiMAX, 3G LTE, 802.11 a/b/g/n, UWB together
in form of intelligent cognitive radio
1143
Generations
Other Tradeoffs:
Rate Rate vs. Coverage
4G Rate vs. Delay
Rate vs. Cost
3G Rate vs. Energy
2G
Mobility
1144
572
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Rule of thumb: the actual capacity (Mbps per channel per sector) in a
multi-cell environment for most wireless
1145
technologies is about 20% to
30% of the peak theoretical data rate.
Wireless evolution to 4G
573
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
4G/IMT-Advanced
International Mobile Telecommunications (IMT)-Advanced Standard
are requirements issued by the ITU-R of the International
Telecommunication Union (ITU) in 2008 for what is marketed as 4G (Or
sometimes as 4,5G mobile phone and Internet access service.
4G provides, in addition to the usual voice and other services of 3G, mobile
broadband Internet access, for example to laptops with wireless modems, to
smartphones, and to other mobile devices.
Potential and current applications include amended mobile web access, IP
telephony, gaming services, high-definition mobile TV, video
conferencing, 3D television, and cloud computing.
4.5G provides better performance than 4G systems, as an interim step
towards deployment of full 5G capability.
The technology includes:
LTE Advanced
MIMO
5G
5G denotes the next major phase of mobile telecommunications standards
beyond the current 4G/IMT-Advanced standards.
5G network requirements could be as:
Spectral efficiency should be significantly enhanced compared to 4G.
Coverage should be improved.
Signaling efficiency enhanced.
Latency should be significantly reduced compared to LTE.
Data rates of several tens of Mb/s should be supported for tens of thousands of
users.
1 Gbit/s to be offered, simultaneously to tens of workers on the same office
floor.
Several hundreds of thousands of simultaneous connections to be supported for
massive sensor deployments.
574
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
5G
Classification
575
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Wireless standards
RAN
IEEE 802.22
WAN
IEEE 802.20
IEEE 802.16e
576
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Frequency Hopping Spread Spectrum (FHSS) - the total frequency band is split into a number of channels.
The broadcast data is spread across the entire frequency band by hopping between the channels in a
pseudorandom fashion.
OFDM - Orthogonal Frequency Division Multiplexing is a multi carrier transmission technique capable of
supporting high speed services whilst still being bandwidth efficient. It achieves this by forcing multiple
1153
sub-carriers together. However, to ensure these adjacent sub-carriers do not cause excessive
interference, they must be orthogonal or 90 to one another.
IEEE 802.11n
Over-the-air (OTA): 200 Mbps; MAC layer (MAC-SAP*): 100Mbps
Microcells, neighborhood area networks (NANs)
PHY
MIMO/multiple antenna techniques
Advanced FEC, (Forward Error Correction)
10, 20 & 40MHz channels widths
Higher order modulation/coding
1154
577
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
578
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
infrastructure S
S
master/slaves:
slaves request permission to
send (to master)
master grants requests
802.15: evolved from
Bluetooth specification
2.4-2.5 GHz radio band
1158
up to 1 Mbps
579
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
0 1 0 1
Frequency
Modulation
2.4 GHz
Communication
Ultrawideband
1 0 1
Impulse
Modulation
(FCC Min=500MHz)
580
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
UWB spectrum
Bluetooth,
802.11b 802.11a
GPS
PCS
-41 dBm/MHz
UWB
Spectrum
1.6 1.9 2.4 3.1 5 10.6
Frequency (GHz)
1162
581
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Low-Rate WPAN
Very low power consumption (no recharge for months or
years!), up to 255 devices
Data rates of 20, 40, 250 kbps
Star clusters or peer-to-peer operation
CSMA-CA channel access
Frequency of operation in ISM bands
Home automation, consumer electronics applications,
RFID/tagging applications (goods supply-chain)
582
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Paging systems
Broad coverage for short messaging
Message broadcast from all base stations
High Tx power (hundreds of watts to kilowatts),
low power pagers
Simple terminals
Optimized for 1-way transmission
Answer-back hard
Overtaken by cellular
obsolete
1166
583
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1168
584
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Licensed
Cell phones, police & fire radio, taxi dispatch, etc.
Unlicensed
All unlicensed bands impose power limits
Industrial, Scientific, Medical (ISM) bands
e.g. (900MHz, 2.4GHz, 5.8GHz)
Unlicensed Personal Communication System (UPCS)
e.g. 1.910-1.920 GHz and 2.390-2.400 GHz
Unlicensed National Information Infrastructure
(UNII) bands
e.g. 5.2GHz
585
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1172
586
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1 2 3 4 5 GHz
1173
802.11/802.16 spectrum
UNII
International International
US Japan ISM
Licensed ISM Licensed Licensed
Licensed
1 2 3 4 5 GHz
1174
587
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Licensed
New Spectrum
2 3 4 5 GHz
1175
588
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
589
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Cells in reality
1179
590
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1181
Issues:
Mobiles traverse a small cell more quickly than a large cell.
Handoffs must be processed more quickly.
Location management becomes more complicated, since there are
more cells within a given area where a mobile may be located.
May need wireless backhaul
Wireless propagation models dont work for small cells.
Microcellular systems are often designed using square or triangular
cell shapes, but these shapes have a large margin of error
1182
591
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Sectoring: Tradeoffs
More antennas.
1184
592
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Global
Satellite
Suburban Urban
In-Building
Picocell
Microcell
Macrocell
Basic Phone
Smart Phone
Laaptop
1185
1186
593
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1187
Handover (Handoff)
Handover :
Cellular system tracks mobile stations in order to maintain their
communication links.
When mobile station goes to neighbor cell, communication link switches
from current cell to the neighbor cell.
Hard handover :
In FDMA or TDMA cellular system, new communication establishes after
breaking current communication at the moment doing handover.
Communication between MS and BS breaks at the moment switching
frequency or time slot.
switching
Cell B Cell A
594
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Soft handover
Soft handover :
In CDMA cellular system, communication does not break even at the
moment doing handover, because switching frequency or time slot is
not required.
Cell B
Cell A
595
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1191
Base Station
Forward link
Reverse link
Mobile Station
1192
596
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Mobile Station
1193
1194
597
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Mobile Station
1195
Transmitter Transmitter
BPF BPF
Receiver F1 F1 Receiver
Synchronous Switches
1196
598
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Multiplexing: Outline
Single link:
Channel partitioning (TDM, FDM, WDM)
vs. Packets/Queuing/Scheduling
Series of links:
Circuit switching vs. packet switching
Statistical Multiplexing (leverage randomness)
Multiplexing gain
Distributed multiplexing (MAC protocols)
Channel partitioning: TDMA, FDMA, CDMA
Randomized protocols: Aloha, Ethernet (CSMA/CD)
Taking turns: distributed round-robin: polling, tokens
1197
Multiplexing: TDM
1198
599
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
p ( )
[ns]
Problem with high rate data
transmission:
multipath delay spread is of the
order of symbol time
inter-symbol-interference (ISI)
1200
600
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Transmitted signal:
Received Signals:
Line-of-sight:
Reflected:
1201
2 Channels Frequency
8 Channels
Frequency
Channels are
narrowband
(flat fading, ISI)
1202
601
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
SISO
Single Input,
Single Output
MISO
Multiple Input,
Single Output
SIMO
Single Input,
Multiple Output
MIMO
Multiple Input,
SDMA Multiple Output
1204
602
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Interference Mitigation
energy reduction
enhanced capacity
improved link budget
Enhanced Rate/Throughput
co-channel streams
increased capacity
increased data rate
1205
Base Station
Forward link
Reverse link
Mobile Station
Mobile Station
Mobile Station Mobile Station
1206
603
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1208
604
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
Channel partitioning
MAC protocols
TDMA: time division multiple access
Access to channel in "rounds"
Each station gets fixed length slot (length = pkt
trans time) in each round
Unused slots go idle
Example: 6-station LAN, 1,3,4 have pkt, slots
2,5,6 idle
Does not statistical multiplexing gains here
1209
TDMA overview
A
B f0
C B A C B A C B A C B A
C
Time
1210
605
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
FDMA overview
C C
f2
B B f1
A A f0
Time
CDMA
spread spectrum
Code B
Code A
B
B
Code A A
A
B C C
B B C
A A A B
A C
B
Time
Sender Receiver
1212
606
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
OFDMA
OFDMA: a mix of FDMA/TDMA: (OFDM modulation)
Sub channels are allocated in the frequency domain,
OFDM symbols allocated in the time domain.
Dynamic scheduling leverages statistical multiplexing gains,
and allows adaptive modulation/coding/power control, user
diversity
t TD M A
T D M A \O F D M A
m
N
1213
FDMA
power
TDMA
power
CDMA
power
1214
607
Lecture notes Telecommunications Engineering II by Jorma Kekalainen
1215
608