Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic...
Transcript of Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic...
![Page 1: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/1.jpg)
Topic 1: Review of Stochastic Processes
Telecommunication Systems Fundamentals
Profs. Javier Ramos & Eduardo Morgado
Academic year 2.013-2.014
![Page 2: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/2.jpg)
Concepts in this Chapter
• Review of Signals models and classification– Examples of actual signals
– Signal modeling
– Signals classification
• Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages
– Autocorrelation
– Independence– Independence
– Stationarity
– Ergodicity
– Cross-correlation
– Power and Energy Spectral Density
Telecommunication Systems Fundamentals
2
Theory classes: 3 sessions (6 hours)
Problems resolution: 1 session (2 hours)
Lab (Matlab): 2 hours
![Page 3: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/3.jpg)
Bibliography
1. Communication Systems Engineering. John. G. Proakis.
Prentice Hall
2. Sistemas de Comunicación. S. Haykin. Wiley
Telecommunication Systems Fundamentals
3
![Page 4: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/4.jpg)
Concepts in this Chapter
• Review of Signals models and classification– Examples of actual signals
– Signal modeling
– Signals classification
• Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages
– Autocorrelation
– Independence– Independence
– Stationarity
– Ergodicity
– Cross-correlation
– Power and Energy Spectral Density
Telecommunication Systems Fundamentals
4
![Page 5: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/5.jpg)
Pure Tone
Noise Free
Telecommunication Systems Fundamentals
Aditive Noise
![Page 6: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/6.jpg)
All frequencies
1 s
Telecommunication Systems Fundamentals
![Page 7: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/7.jpg)
Why modeling Signals?
• To answer the following questions:
– What information does the signal contain? How is the info coded into the signal? How much info does the signal contain?
– How does the channel affect the transmitted signal?
– How is the telecommunication system designed?– How is the telecommunication system designed?
• We describe signals by their mathematical model –
measurable characteristics of the signal
Telecommunication Systems Fundamentals
![Page 8: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/8.jpg)
How would you describe them?
Telecommunication Systems Fundamentals
![Page 9: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/9.jpg)
Signal Modeling
• In a point-by-point description, the value of the signal at each time instant is stored in a look-up-table
t x(t)
… …
0 7
• The point-by-point description is valid for any signal (assuming the sampling rate is fast enough) and contains all the information within the signal, but “seeing” the information is not evident
0 7
1 2
… …
Telecommunication Systems Fundamentals
![Page 10: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/10.jpg)
Signal Modeling
• Some signals can be modeled by a mathematical expression that provides its amplitude as function of time
x(t) = An
n =0
∞
∑ cos(nω0t)
• This type of signals are named “Deterministic” because their lack of randomness
– Only few signals in telecommunications systems can be modeled as simple as
this
Telecommunication Systems Fundamentals
![Page 11: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/11.jpg)
Signal Modeling
• We can briefly describe a signal by some of its characteristics
– Mean value
– Mean squared value (power)
– Energy
– Standard deviation
– Autocorrelation
– …– …
• It is a universal procedure (usable for any kind of signal), and it gives some criteria to classify signals. However, it does not describe the signal completely (univocally).
Telecommunication Systems Fundamentals
![Page 12: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/12.jpg)
How would you describe them?
Telecommunication Systems Fundamentals
![Page 13: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/13.jpg)
Mean Value
• For time-discrete signals, mean value is defined as:
• For time-continuous signals, mean value is defined as:
x[n] = limN →∞
1
2N +1x[n]
n=−N
N
∑
• For time-continuous signals, mean value is defined as:
x(t) = limT →∞
1
2Tx(t)dt
−T
T
∫
Telecommunication Systems Fundamentals
![Page 14: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/14.jpg)
Mean Value
-0.5
0
0.5
1
1.5
2
0
1
2
3
4
5
5
10
15
20
25
0 5 10 15 20 25 30 35 40-1
-0.5
10 20 30 40 50 60 70 80-1
0
5 10 15 20 25 30 35 40
0
2)(2 =tx
Telecommunication Systems Fundamentals
0)(1 =tx ∞=)(3 tx
![Page 15: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/15.jpg)
Energy
• A decisive classification of signals is related to its
energy and power: finite energy, or power defined. For
finite energy signals, it is defined
– For discrete signals:
Ex
= x[n]2
∞
∑
– For continuous signals:
Ex
= x(t)2dt
−∞
∞
∫
Ex
= x[n]n=−∞
∑
Telecommunication Systems Fundamentals
![Page 16: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/16.jpg)
Energy
0.5
1
1.5
2
2
3
4
5
10
15
20
25
0 5 10 15 20 25 30 35 40-1
-0.5
0
10 20 30 40 50 60 70 80-1
0
1
5 10 15 20 25 30 35 40
0
5
10
10=1
xE ∞=
2x
E ∞=3x
E
Telecommunication Systems Fundamentals
![Page 17: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/17.jpg)
Average Power
• The average power of discrete signals is defined as:
• While the average power of continuous signal is
Px
= limN →∞
1
2N +1x[n]
2
n=−N
N
∑
• While the average power of continuous signal is
defined as:
Px
= limT →∞
1
2Tx(t)
2dt
−T
T
∫
Telecommunication Systems Fundamentals
![Page 18: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/18.jpg)
Average Power
0
0.5
1
1.5
2
1
2
3
4
5
10
15
20
25
0 5 10 15 20 25 30 35 40-1
-0.5
10 20 30 40 50 60 70 80-1
0
5 10 15 20 25 30 35 40
0
5
0=1
xP 5=
2x
P
Telecommunication Systems Fundamentals
Goes to zero when
the length of the
samples increases
∞=3x
P
![Page 19: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/19.jpg)
A signal classification
Signal
Finite
Energy
Finite
AverageUnbounded
AverageEnergy
(Energy Defined)
Average
Power
(Power Defined)
Average
Power
0 ≤ Ex
< ∞
Px
= 0
Ex
= ∞
0 < Px
< ∞
Ex
= ∞
Px
= ∞
Telecommunication Systems Fundamentals
![Page 20: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/20.jpg)
Energy/Power signal classification
-0.5
0
0.5
1
1.5
2
0
1
2
3
4
5
0
5
10
15
20
25
0 5 10 15 20 25 30 35 40-1
10 20 30 40 50 60 70 80-1
5 10 15 20 25 30 35 40
0
Telecommunication Systems Fundamentals
Finite
Energy
(Energy Defined)
Finite
Average
Power
(Power Defined)
Unbounded
Average
Power
(assuming it increases
indefinitely)
![Page 21: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/21.jpg)
Energy/Power signal classification
Finite Energy
(Energy Defined)
Finite Average Power
Telecommunication Systems Fundamentals
Finite Average Power
(Power Defined)
Unbounded Average Power
![Page 22: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/22.jpg)
Homework
• Compute: i) Average Value; ii) Energy and ii) Average Power of the two following signals:
x (t) = Aej 2πft
x1(t) = Acos(2πft)
x2(t) = Aej 2πft
Telecommunication Systems Fundamentals
![Page 23: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/23.jpg)
Classifying Signals: A Taxonomy
� Continuous / Discrete
� Analog / Digital
� Deterministic / Stochastic (random signals)
� Deterministic:
� Energy Defined (time limited)
� Power Defined
� Periodic / Non periodic� Periodic / Non periodic
� Stochastic
� Stationary
� Ergodic / Non-Ergodic
� Non-Stationary
� Other classifications
� Real valued / Complex
� Even / Odd
� Hermitical / Non-Hermitical
23
Telecommunication Systems Fundamentals
![Page 24: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/24.jpg)
Concepts in this Chapter
• Review of Signals models and classification– Examples of actual signals
– Signal modeling
– Signals classification
• Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages
– Autocorrelation
– Independence– Independence
– Stationarity
– Ergodicity
– Cross-correlation
– Power and Energy Spectral Density
Telecommunication Systems Fundamentals
24
![Page 25: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/25.jpg)
Recall: Time Averaging and Expected Value
x(t) = lim
T → ∞
1
2Tx(t)dt
−T
T
∫
E(X) = xfX (x)dx
−∞
∞
∫
Telecommunication Systems Fundamentals
![Page 26: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/26.jpg)
Homework: Time Averaging and Expected Value
• Generate a Random Variable uniformly distributed
between 0 and 1
– X ~U(0,1)
– Pdf - f(x) = 1, for 0 < x < 1, and 0 otherwise.
1. Run a simulation (Matlab) of 10.000 samples of U(0,1)1. Run a simulation (Matlab) of 10.000 samples of U(0,1)
2. Compute the average value of the 10.000 samples
3. Analytically calculate expected value of U(0,1)
4. Compare values obtained in 2 and 3.
Telecommunication Systems Fundamentals
![Page 27: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/27.jpg)
Recall: Time Averaging and Expected Value
x
2 (t) = limT → ∞
1
2Tx
2 (t)dt−T
T
∫
E(X
2) = x2
fX (x)dx−∞
∞
∫
Telecommunication Systems Fundamentals
![Page 28: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/28.jpg)
Homework: Time Averaging and Expected Value
• Generate a Random Variable uniformly distributed
between 0 and 1
– X ~U(0,1)
– Pdf - f(x) = 1, for 0 < x < 1, and 0 otherwise.
1. Run a simulation (Matlab) of 10.000 samples of U(0,1)1. Run a simulation (Matlab) of 10.000 samples of U(0,1)
2. Compute the average power of the 10.000 samples
3. Analytically calculate second moment of U(0,1)
4. Compare values obtained in 2 and 3.
Telecommunication Systems Fundamentals
![Page 29: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/29.jpg)
Why statistical modeling is useful?
1. Characterizing a stochastic process would require the specification of the signal at every single instant
2. Most cases we do not know the signal a priori
3. We get the whole signal in very rare occasions
Statistical model to:Statistical model to:
… sumarize the description of a signal behaviour
… describe the whole signal from a finite time interval
… describe sets of signals
Telecommunication Systems Fundamentals
![Page 30: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/30.jpg)
0 0 . 5 1 1 . 5 2 2 . 5 3- 1
- 0 . 8
- 0 . 6
- 0 . 4
- 0 . 2
0
0 . 2
0 . 4
0 . 6
0 . 8
1
t i m e
mag
nit
ud
e
0 0 . 5 1 1 . 5 2 2 . 5 30
2
4
6
8
1 0
1 2
1 4
t i m e
mag
nit
ud
e
How do you describe them?
0 0 . 5 1 1 . 5 2 2 . 5 3- 2
- 1 . 5
- 1
- 0 . 5
0
0 . 5
1
1 . 5
2
t i m e
ma
gn
itu
de
0 0 . 5 1 1 . 5 2 2 . 5 30
2
4
6
8
1 0
1 2
1 4
t i m e
ma
gn
itu
de
Telecommunication Systems Fundamentals
Why Statistical Modeling is Useful?
![Page 31: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/31.jpg)
What’s next?
0 1 2 3 4 5 60
2
4
6
8
1 0
1 2
1 4
t i m e
ma
gn
itu
de
t i m e
0 1 2 3 4 5 6- 2
- 1 . 5
- 1
- 0 . 5
0
0 . 5
1
1 . 5
2
t i m e
ma
gn
itu
de
Telecommunication Systems Fundamentals
![Page 32: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/32.jpg)
0 1 2 3 4 5 60
2
4
6
8
1 0
1 2
1 4
t i m e
ma
gn
itu
de
What’s next?
0 1 2 3 4 5 6- 2
- 1 . 5
- 1
- 0 . 5
0
0 . 5
1
1 . 5
2
t i m e
ma
gn
itu
de
t i m e
Telecommunication Systems Fundamentals
![Page 33: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/33.jpg)
Review of the Concept of Stochastic Process
• Definition 1: a SP can be seen as series of
Random Variables; or it can be also seen
as a RV that is time-variant
• Definition 2: a SP can be seen as a set of
time-variant signals, each one with its
probability of happening (imagine a bag
with all the possible signals and you get
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
t i m e
magn
itud
e
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
t i m e
magn
itud
e
3
4
5
6
magn
itud
e
X1(t) � P1
X2(t) � P2
X3(t) � P3
X4(t) � P4with all the possible signals and you get
one of them).
• Note: we talk about “time” when referring to
the independent variable, but it could be
different:– Example: thickness of bar as function of its length X(t)
0 2 4 6 8 1 0 1 20
1
2
t i m e
Telecommunication Systems Fundamentals
![Page 34: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/34.jpg)
Stochastic Process Model
• To fully characterize an SP a probability measurement of each possible realization has to be provided.
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
t i m e
ma
gn
itu
de
3
4
5
6
P1
P2
• In other words, we should be able to tell how likely is any given observation (realization) to happen
0 2 4 6 8 1 0 1 20
1
2
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
t i m e
ma
gn
itu
de
.
.
.
P3
Telecommunication Systems Fundamentals
![Page 35: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/35.jpg)
Stochastic Process Model
• A complete description of a SP, X(t), requires the definition of the sequence (X(t1), …, X(tk)) for any value of k and any value of the k-tuple (t1, …, tn).
t1 t2 t3 t4 t5
Telecommunication Systems Fundamentals
![Page 36: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/36.jpg)
Stochastic Process Model
• In general practice, we will not look for a complete
description of the SP, but we will define by two main
aspects:
– Amplitude distribution
– Autocorrelation, which contains the time variation description (statistical relationship betwen two instants of the signal)
• Autocorrelation can be expressed also as Power
Spectral Density – the Fourier Transform of the
Autocorrelation
• Later, we can analyse the impact of a linear channel on
the SP, i.e. the impact of the channel on the amplitude
distribution and on the autocorrelation
Telecommunication Systems Fundamentals
![Page 37: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/37.jpg)
Summarizing main concepts of SP
• A SP is a mechanism that generates time-variant amplitudes – a signal. Each of the signal produced by a SP is called “realization”
• The SP model also applies to each realization. In other words, a model for a SP models also every possible realization of it.
• We will model SP by their amplitude distribution and autocorrelation. • We will model SP by their amplitude distribution and autocorrelation. Amplitude distribution models the realization values at a given time, and autocorrelation models the time variation of the SP.
• By computing Fourier Transform of autocorrelation we get the Power Spectral Density – the information of the amount of energy contained in each frequency – the spectrum
• Noise in telecommunications is modeled as a SP
Telecommunication Systems Fundamentals
![Page 38: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/38.jpg)
Amplitude Distribution
• For each instant of time of the SP (sample), its amplitude is a Random Variable following a Probability Density Function (pdf)
• Amplitude Distribution is modeled by its pdf. It can be modeled both by its amplitude (two values for complex signals) or by its power. Thus, a pdf of the signal amplitude or a pdf of the signal power should be provided
5
6
7
Most frequent
0 2 0 4 0 6 0 8 0 1 0 0- 1
0
1
2
3
4
5
t i m e
mag
nit
ud
e
0 1 2 3 4 5 6 70
5
10
15
20
25
30
35
40
magnitude
# o
ccu
rre
nc
es
Most frequent
value Less frequent
value
Telecommunication Systems Fundamentals
![Page 39: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/39.jpg)
Amplitude Distribution
-1.5 -1 -0.5 0 0.5 1 1.5magnitude
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.80
1000
2000
3000
4000
5000
6000
7000
8000
-1 -0.5 0 0.5 1magnitude
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.30
0.5
1
1.5
2
2.5x 10
4
Telecommunication Systems Fundamentals
![Page 40: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/40.jpg)
Homework
• Specify the magnitudes on each axes of the above graph
– A) If we interpret the plot as a Gaussian-shaped signal– A) If we interpret the plot as a Gaussian-shaped signal
– B) If we interpret the plot as the pdf of the voltage of a noisy signal
• What is the mean value for each case?
Telecommunication Systems Fundamentals
![Page 41: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/41.jpg)
Mean value: time domain and statistical approach
• Time average
- 1
- 0 .5
0
0 .5
1
ma
gn
itu
de
• Statistical mean value
- 2 - 1 .5 - 1 - 0 .5 0 0 .5 1 1 .5 2- 1
t im e
x(t) = lim
T → ∞
1
2Tx(t)dt
−T
T
∫
x
2 (t) = limT → ∞
1
2Tx
2 (t)dt−T
T
∫
E (X ) = xf X (x)dx
−∞
∞
∫
E(X
2) = x2fX (x)dx
−∞
∞
∫
-1.5 -1 -0.5 0 0.5 1 1.5magnitude
Telecommunication Systems Fundamentals
![Page 42: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/42.jpg)
Numerical Example
• Time averagex[n] = 5 6 2 1 2 3 2 6 1 4 1 6 2 4 6
3 1 2 3 3 2 6 5 6 3 6 1 4 5 2 6 5 5 4 5 2 2 2 4 5 2 3 6 4 6 2 5 5 2 3 3 4 5 1 5 6 3 3 4 5 5 5 3 3 6 4 6 2 4 4 6 1 4 4 1 6 ...
[ ] 7,3][1 1
0
== ∑−
=
N
n
nxN
nx
Roll a Dice
• Statiscal average
E X( ) = nf x n[ ]n =1
N
∑ =
= 1⋅1
6+ 2⋅
1
6+ 3⋅
1
6+ 4 ⋅
1
6+ 5⋅
1
6+ 6⋅
1
6= 3.5
3,7 for 100 samples, for 100.000 � 3,4997
Telecommunication Systems Fundamentals
![Page 43: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/43.jpg)
The Correlogram
John Henry Poynting(1852 - 1914)
Telecommunication Systems Fundamentals
![Page 44: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/44.jpg)
Amplitude distribution is not enough to describe a time-variant SP
0 10 20 30 40 50 600
1
2
3
4
5
6
7
8
tim e
ma
gn
itu
de
0 10 20 30 40 50 600
1
2
3
4
5
6
7
8
axis
mag
nit
ud
e
- 1 0 - 5 0 5 1 00
0 . 0 2
0 . 0 4
0 . 0 6
0 . 0 8
0 . 1
0 . 1 2
0 . 1 4
0 . 1 6
0 . 1 8
m a g n i t u d e
Telecommunication Systems Fundamentals
![Page 45: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/45.jpg)
Autocorrelation
• Many power-defined signals (time-unbounded) exhibit repetition patterns. Although such signals are not periodic, they have some periodicity on their amplitude distribution. They are quasi-periodic
• How can we study such signals?
• An approach to analyze quasi-periodic patterns is to check the likeness between the signal and a delayed version of itself
• The autocorrelation function describes the likeness of a signal with a delayed version of itself. Therefore, we can identify quasi-periodic patterns by computing the signal autocorrelation
Telecommunication Systems Fundamentals
![Page 46: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/46.jpg)
Autocorrelation
• To measure likeness between a given signal and a delayed version of itself we use the inner product of both signals. So, autocorrelation is defined that way.
• The likeness measurement (inner product) is defined in different way for power-defined and energy-defined signals
• Calculation of inner product depends on the available information of the signal
– If time description of realizations is available, we can compute inner product
as time average
– If statistical information is available, inner product will be computed as
statistical average
Telecommunication Systems Fundamentals
![Page 47: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/47.jpg)
Example 1: Energy-Defined Signal
1
2
-10 -5 0 5 10
1
Telecommunication Systems Fundamentals
![Page 48: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/48.jpg)
Example 1: Energy-Defined Signal
…
x(-5)x(-5) = 1
x(-4)x(-4) = 0
x(-3)x(-3) = 1
x(-2)x(-2) = 0
x(-1)x(-1) = 1
0.5
1
1.5
x(-1)x(-1) = 1
x(0)x(0) = 0
x(1)x(1) = 1
x(2)x(2) = 0
x(3)x(3) = 1
…
Σx(n)x(n) = 6
-10 -5 0 5 100
-10 -5 0 5 100
0.5
1
1.5
Telecommunication Systems Fundamentals
![Page 49: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/49.jpg)
Example 1: Energy-Defined Signal
…
x(-5)x(-4) = 0
x(-4)x(-3) = 0
x(-3)x(-2) = 0
x(-2)x(-1) = 0
x(-1)x(0) = 0-10 -5 0 5 100
0.5
1
1.5
x(-1)x(0) = 0
x(0)x(1) = 0
x(1)x(2) = 0
x(2)x(3) = 0
x(3)x(4) = 0
…
Σx(n)x(n+1) = 0
-10 -5 0 5 10
-10 -5 0 5 100
0.5
1
1.5
Telecommunication Systems Fundamentals
![Page 50: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/50.jpg)
Example 1: Energy-Defined Signal
…
x(-5)x(-3) = 1
x(-4)x(-2) = 0
x(-3)x(-1) = 1
x(-2)x(0) = 0
x(-1)x(1) = 1-10 -5 0 5 100
0.5
1
1.5
x(-1)x(1) = 1
x(0)x(2) = 0
x(1)x(3) = 1
x(2)x(4) = 0
x(3)x(5) = 1
…
Σx(n)x(n+2) = 5
-10 -5 0 5 10
-10 -5 0 5 100
0.5
1
1.5
Telecommunication Systems Fundamentals
![Page 51: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/51.jpg)
Example 1: Energy-Defined Signal
…
x(-5)x(-2) = 0
x(-4)x(-1) = 0
x(-3)x(0) = 0
x(-2)x(1) = 0
x(-1)x(2) = 0-10 -5 0 5 100
0.5
1
1.5
x(0)x(3) = 0
x(1)x(4) = 0
x(2)x(5) = 0
x(3)x(6) = 0
…
Σx(n)x(n+3) = 0
-10 -5 0 5 100
0.5
1
1.5
Telecommunication Systems Fundamentals
![Page 52: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/52.jpg)
Example 1: Energy-Defined Signal
…
x(-5)x(-1) = 1
x(-4)x(0) = 0
x(-3)x(1) = 1
x(-2)x(2) = 0
x(-1)x(3) = 1-10 -5 0 5 100
0.5
1
1.5
x(-1)x(3) = 1
x(0)x(4) = 0
x(1)x(5) = 1
x(2)x(6) = 0
x(3)x(7) = 0
…
Σx(n)x(n+3) = 4
-10 -5 0 5 100
0.5
1
1.5
Telecommunication Systems Fundamentals
![Page 53: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/53.jpg)
Example 1: Energy-Defined Signal
2
4
5
6
7
-10 -5 0 5 10
1
-10 -5 0 5 100
1
2
3
x(n) Rx(k)=Σx(n)x(n+k)
Telecommunication Systems Fundamentals
![Page 54: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/54.jpg)
1
2
Example 2: Energy-Defined Signal
-10 -5 0 5 10-2
-1
0
Telecommunication Systems Fundamentals
![Page 55: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/55.jpg)
0
5
10
0
1
2
Example 2: Energy-Defined Signal
-10 -5 0 5 10
-10
-5
-10 -5 0 5 10-2
-1
x(n) Rx(k)=Σx(n)x(n+k)
Telecommunication Systems Fundamentals
![Page 56: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/56.jpg)
Autocorrelation for Energy-Defined Signals
• If x[n] is a discrete signal energy-defined, its autocorrelation function Rx[k] is defined as:
Rx[k] = x[n]
n=−∞
∞
∑ x[n − k]
= x[k]∗ x[−k]
• If x(t) is a continuous-time signal energy-defined, its autocorrelation function Rx(t) is defined as:
Rx(τ ) = x(t)x(t − τ )dt
−∞
∞
∫= x(τ ) ∗ x(−τ )
Telecommunication Systems Fundamentals
![Page 57: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/57.jpg)
0
1
2
Example 3: Power-Defined Signal
-0.5
0
0.5
1
1.5
-20 -15 -10 -5 0 5 10 15-2
-1
x(n)
-20 -15 -10 -5 0 5 10 15 20-1.5
-1
-0.5
Telecommunication Systems Fundamentals
][kRx
![Page 58: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/58.jpg)
0.5
1
1.5
0
1
2
Example 4: Power-Defined Signal
-20 -15 -10 -5 0 5 10 15 20-1
-0.5
0
-20 -15 -10 -5 0 5 10 15 20-2
-1
x(n)
Telecommunication Systems Fundamentals
][kRx
![Page 59: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/59.jpg)
Autocorrelation for Power-Defined Signals
• If x[n] is a discrete signal power-defined, its autocorrelation function Rx[k] is defined as:
Rx[k] = lim
N → ∞
1
2N +1x[n]
n =−N
N
∑ x[n − k]
• If x(t) is a continuous-time signal power-defined, its autocorrelation function Rx(t) is defined as:
Telecommunication Systems Fundamentals
Rx(τ) = lim
T →∞
1
2Tx(t)x(t − τ)dt
−T
T
∫
![Page 60: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/60.jpg)
Homework
• Compute the autocorrelation of: x(t) = Acos 2πft( )
Telecommunication Systems Fundamentals
![Page 61: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/61.jpg)
Autocorrelation Summary
• The autocorrelation function is a measurement of the self-likeness of a signal – in other words, the autocorrelation gives information about the likeness of a signal with itself but delayed
• Therefore, the autocorrelation function summarizes the time behavior of a signalbehavior of a signal
• Mathematically, the autocorrelation function is a projection (inner product) of a signal against a delayed version of itself with all possible values of delay. Consequently, the maximum of the autocorrelation function is at delay equal to zero – the likeness of a signal with itself – its power or energy
Telecommunication Systems Fundamentals
![Page 62: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/62.jpg)
Autocorrelation Properties
• Autocorrelation definition is different for energy-defined and power-define signals.
• In both cases, autocorrelation measures signal likeness to delayed version of itself, referring this likeness to its maximum value which happens at delay equal to zero.equal to zero.
• For energy-defined signals
• And for power-defined signals
xER =)0(x
xPR =)0(x
Telecommunication Systems Fundamentals
![Page 63: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/63.jpg)
Autocorrelation Properties
• Autocorrelation function satisfies:
• And for the particular case of periodic signals, it holds that:
Rx(τ ) ≤ R
x(0)
-20 -15 -10 -5 0 5 10 15 20-1
-0.5
0
0.5
1
1.5
-20 -15 -10 -5 0 5 10 15 20-2
-1
0
1
2
• And for the particular case of periodic signals, it holds that:
where L is any integer number and T is the signal period
Rx(L ⋅ T ) = R
x(0)
-20 -15 -10 -5 0 5 10 15 20-1.5
-1
-0.5
0
0.5
1
1.5
-20 -15 -10 -5 0 5 10 15-2
-1
0
1
2
Telecommunication Systems Fundamentals
![Page 64: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/64.jpg)
Autocorrelation Properties
– Symmetry: RX(τ) is and even signal: RX(τ) = RX(- τ).
– Maximum: RX(τ) maximum is for τ = 0 (and coincides with energy/power of the signal), | RX(τ) | ≤ RX(0).
– Periodicity: if for a give value of T, it holds that RX(T) = RX(0), – Periodicity: if for a give value of T, it holds that RX(T) = RX(0), then it also holds that RX(kT+ τ) = RX(0+ τ) for any integer value of k.
– Integrability: the autocorrelation function of any signal (except the periodic signals) can be integrated – i.e. it is a energy-defined signal itself.
Telecommunication Systems Fundamentals
![Page 65: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/65.jpg)
Autocorrelation of Stochastic Processes
• Recall the two viewpoints for SP:
– A set of signals with common properties, although signals itself are not identical point-by-point
– A physical mechanism that generates sets of signals according to a stochastic pattern
• In any case, to describe an SP the set of signals has to • In any case, to describe an SP the set of signals has to
be described, but a single signal can not be described.
• So, how can we define the autocorrelation of a SP?
Telecommunication Systems Fundamentals
![Page 66: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/66.jpg)
Intuition
• Let’s assume, as starting point, that
we have a set of signals and we
select one of them by a random
mechanism.
• The autocorrelation of one of these
realizations, x1(t), can be computed
as its time average using next
expression:
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
t i m e
magn
itud
e
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
0 2 4 6 8 1 0 1 20
1
2
3
4
5
6
t i m e
magn
itud
e
1
2
3
4
5
6
magn
itud
e
X1(t) � Rxx,1X2(t) � Rxx,2
X3(t) � Rxx,3
X4(t) � Rxx,4
expression:
R
xx,1(τ) = limT →∞
1
2Tx1(t)x1(t −τ)dt
−T
T
∫ X(t)
0 2 4 6 8 1 0 1 20
t i m e
Telecommunication Systems Fundamentals
![Page 67: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/67.jpg)
Intuition
• But if we compute autocorrelation in such a way, we are not
computing the autocorrelation of the SP, but the one of a
particular signal (realization)
• So, in order to compute the autocorrelation of the SP, we
should average over all possible realizations:should average over all possible realizations:
Rx(τ) = E lim
T →∞
1
2Tx(t)x(t − τ)dt
−T
T
∫
= limT →∞
E1
2Tx(t)x(t − τ)dt
−T
T
∫
= limT →∞
1
2TE x(t)x(t − τ)[ ]dt
−T
T
∫
Telecommunication Systems Fundamentals
![Page 68: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/68.jpg)
Intuition ���� Definition
• If the expected value does not depend on the time, then:
• The autocorrelation function for a SP is noted as RX(t1,t2), and its definition is
R (t ,t ) = E [X(t ) X(t )]:
Rx(τ) = E x(t)x(t − τ)[ ]lim
T →∞
1
2Tdt
−T
T
∫= E x(t)x(t − τ)[ ]
RX(t1,t2) = E [X(t1) X(t2)]:
• Which is a measurement of the likeness between Random Variables obtained in two instants of the SP, X(t1) and X(t2). i.e. RX(t1,t2) is a measurement of the likeness (variation) of the signal in two different instants of time t1 and t2.
RX (t1, t2) = E X(t1)X(t2)[ ]
= x1−∞
∞
∫ x2 fX (t1 ),X (t2 )(x1, x2)dx1dx2−∞
∞
∫
Telecommunication Systems Fundamentals
![Page 69: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/69.jpg)
Example: Rxx based on one SP realization
x(t) Rxx(t)
Onerealization
of SP 1
Onerealization
of SP 2
Telecommunication Systems Fundamentals
of SP 2
Onerealization
of SP 3
Autocorrelations obtained usingonly the single realization
SP 3 more correlated than SP 2, which is more correlated than SP 1
(uncorrelated)
![Page 70: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/70.jpg)
Example: Rxx averaging over “all” (many) SP realizations
SP 3 more correlated thanSP 2, which is
Telecommunication Systems Fundamentals
SP 2, which ismore correlated
than SP 1 (uncorrelated)
![Page 71: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/71.jpg)
Uncorrelated Processes
• A SP is uncorrelated if RX(t1,t2) = 0, for any t1 and t2such that t1 ≠ t2.
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
– Intuitively, there is not “likeness” between samples at different times of the process
- 5 0 5
x 1 0- 3
0
1
2
3
4
5
6
7
8
Telecommunication Systems Fundamentals
![Page 72: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/72.jpg)
Noise in Telecommunications often is Uncorrelated
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
- 5 0 5
x 1 0- 3
0
1
2
3
4
5
6
7
8
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
x 1 0
- 5 0 5
x 1 0- 3
0
1
2
3
4
5
6
7
8
9
Telecommunication Systems Fundamentals
![Page 73: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/73.jpg)
Sine on/off wave (cricket)
270 ms
0,2 ms45 ms
Telecommunication Systems Fundamentals
![Page 74: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/74.jpg)
0,2 ms
Sine on/off wave
270 ms
0,2 ms45 ms
Telecommunication Systems Fundamentals
![Page 75: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/75.jpg)
Summarizing
• Autocorrelation is a measurement of the likeness
between a signal and a delayed version of itself
• Each realization of a SP may be quite different and
computing any statistic on it will mislead to totally
incorrect information. Autocorrelation has to average incorrect information. Autocorrelation has to average
all (many) realizations, i.e. compute the expected value
• Peak values in the autocorrelation function hints about
repetition patterns
Telecommunication Systems Fundamentals
![Page 76: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/76.jpg)
Statistical Independence
• Two scenarios:
– Independence of the samples of a signal
– Independence of two signals
• Intuitively, two samples are independent if the mechanisms that • Intuitively, two samples are independent if the mechanisms that generates them are also independent
• Formally, two Random Variables, X1 and X2, are independent if, and only if,
Telecommunication Systems Fundamentals
)()(),( 2121 2121xfxfxxf XXXX ⋅=
![Page 77: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/77.jpg)
Intuition for Independent Samples
Rolling a dice
3, 1, 5, 1, 4, 2, 2, 1, 6, 4, 3Independent:
Dependent:
4, 6, 6, 5, 6, 4, 3, 7, 10, 7, …
3, 1, 5, 1, 4, 2, 2, 1, 6, 4, 3, …
Telecommunication Systems Fundamentals
![Page 78: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/78.jpg)
Intuition for Independent Signals
3, 1, 5, 1, 4, 2, 2, 1, 6, 4, 3
1, 1, 6, 4, 3, 1, 2, 4, 5, 5, 2
Independent?
3, 1, 5, 1, 4, 2, 2, 1, 6, 4, 3
1, 1, 6, 4, 3, 1, 2, 4, 5, 5, 2
4, 2, 11, 5, 7, 3, 4, 5, 11, 9, 5
Telecommunication Systems Fundamentals
Dependent:
Independent:
Independent?
![Page 79: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/79.jpg)
SP and Stationarity
• An SP is stationary if its statistics do not depend on
time
0
1
2
3
4
5
6
7
8
mag
nit
ud
e
Stationary
0 1 0 2 0 3 0 4 0 5 0 6 00
t i m e
0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 00
2
4
6
8
1 0
1 2
1 4
1 6
1 8
2 0
t i m e
mag
nit
ud
e Non-Stationary
Telecommunication Systems Fundamentals
![Page 80: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/80.jpg)
Stationary Stochastic Processes
• When arriving to practical SP, two types of stationarity can be defined:– Strict (Sense) Stationary Processes (SSP). The pdf of the process does not
change with time delay ∆
– Wide Sense Stationary Processes (WSSP), those that satisfy the two following restrictions:
• The mean value of the process, E{X(t)}, does not varies with time
),...,,(),...,,( 2121 ∆+∆+∆+= nXnX tttftttf
• The mean value of the process, E{X(t)}, does not varies with time
• The autocorrelation RX(t1,t2) depends only on the time difference τ = t1 - t2. Thus, for WSSP we compute the autocorrelation as RX(τ).
• An SSP is also WSSP, but the opposite is not true
Telecommunication Systems Fundamentals
Rx(τ) = E x(t)x(t −τ)[ ]
RX (t1, t2) = E X(t1)X(t2)[ ]
![Page 81: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/81.jpg)
SP and Ergodicity
• A Stochastic Process is Ergodic if its statistics can be
computed as time average of one of its realizations
limT →∞
1
2Tg(x(t))dt = E(g(X(t)))
−T
T
∫
• Ergodicity implies stationarity. Any ergodic process is
stationary
• However, stationarity does not imply ergodicity
• Example: roll a dice infinite times
Telecommunication Systems Fundamentals
![Page 82: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/82.jpg)
Example 7
• The following SP is not stationary, and therefore it is
not ergodic either.
0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 00
2
4
6
8
1 0
1 2
1 4
1 6
1 8
2 0
t i m e
mag
nit
ud
e
Telecommunication Systems Fundamentals
![Page 83: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/83.jpg)
Example 8
• The following Strict Sense Stationary Process is not
Ergodic. Why?
1 0 2 0 3 0 4 0 5 0 6 00
1
2
3
4
5
6
7
1 0 2 0 3 0 4 0 5 0 6 00
1
2
3
4
5
6
7
Telecommunication Systems Fundamentals
![Page 84: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/84.jpg)
Time Averaging and Statistical Expected Value
• Time average and expected value returns the same
value only for ergodic SP
• As example, if we assume that human voice signal
corresponds to a ergodic SP, then we can estimate
statistics from time averaging only one recorded piece statistics from time averaging only one recorded piece
of voice, and assume that those values coincide with
the statistics of any voice signal
Telecommunication Systems Fundamentals
![Page 85: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/85.jpg)
x(t) = lim
T → ∞
1
2Tx(t)dt
−T
T
∫
E(X) = xfX (x)dx
−∞
∞
∫
Time Averaging and Statistical Expected Value
x
2 (t) = limT → ∞
1
2Tx
2 (t)dt−T
T
∫
E(X
2) = x2fX (x)dx
−∞
∞
∫
R
x(τ) = lim
T → ∞
1
2Tx(t)x(t − τ)dt
−T
T
∫
RX (τ) = E(X(t)X (t −τ))
Telecommunication Systems Fundamentals
![Page 86: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/86.jpg)
SP classification
Non-Stationary (both power and
Stationary (power
defined signals)
(both power and
energy defined
signals)Ergodic (power
defined signals)
Telecommunication Systems Fundamentals
![Page 87: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/87.jpg)
What is the practical meaning of Stationarity?
• We can characterize one realization from a segment of it
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
- 5 0 5
x 1 0- 3
0
1
2
3
4
5
6
7
8
-5 -4 -3 -2 -1 0 1 2 3 40
1000
2000
3000
4000
5000
6000
Telecommunication Systems Fundamentals
![Page 88: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/88.jpg)
What is the practical meaning of Ergodicity?
• We can characterize the SP from a realization of it
2000
3000
4000
5000
6000
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
- 5 0 5
x 1 0- 3
0
1
2
3
4
5
6
7
8
-5 -4 -3 -2 -1 0 1 2 3 40
1000
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
Telecommunication Systems Fundamentals
![Page 89: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/89.jpg)
Cross-Correlation
• If autocorrelation is a measurement of how a signal
looks like itself delayed; the cross-correlation function
provides information about likeness of two signals
(delaying one respect the other)
• Depending of the type of signals, cross-correlation can • Depending of the type of signals, cross-correlation can
be defined for:
– Cross-Correlation of Energy Defined signals
– Cross-Correlation of Power Defined signals
– Cross-Correlation of one Energy Defined signal and one Power Defined signal
Telecommunication Systems Fundamentals
![Page 90: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/90.jpg)
Cross-Correlation of two Energy Defined Signals
• Let x[n] and y[n] be two energy defined discrete signals (or one Energy defined and the other Power defined). Their cross-correlation function is defined as:
Rxy[k] = x[n]
n=−∞
∞
∑ y[n − k]
= x[k]∗ y[−k]
• If x(t) and y(t) are two energy defined continuous signals (or one Energy defined and the other Power defined). Their cross-correlation function is defined as:
= x[k]∗ y[−k]
Rxy (τ) = x(t)y(t − τ )dt
−∞
∞
∫ = x(τ) ∗ y(−τ )
Telecommunication Systems Fundamentals
![Page 91: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/91.jpg)
Example 1
x[n]
y[n]
-10 -5 0 5 100
0.5
1
1.5
0
5
Rxy[k]
Ryx[k]
-10 -5 0 5 100
-10 -5 0 5 100
5
10
-10 -5 0 5 100
5
10
Telecommunication Systems Fundamentals
![Page 92: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/92.jpg)
Example 2
x[n]
y[n]
-10 -5 0 5 100
0.5
1
1.5
-10 -5 0 5 100
0.5
1
1.5
Rxy[k]
Ryx[k]
-10 -5 0 5 100
-10 -5 0 5 100
5
-10 -5 0 5 100
5
Telecommunication Systems Fundamentals
![Page 93: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/93.jpg)
Cross-Correlation of two Power Defined Signals
• Let x[n] and y[n] be two power defined discrete signals.
Their cross-correlation function, Rxy[k], is defined as:
• Let x(t) and y(t) be two power defined continuous
Rxy[k] = lim
n →∞
1
2n +1x[n]
n=−∞
∞
∑ y[n − k]
• Let x(t) and y(t) be two power defined continuous
signals. Their cross-correlation function, Rxy(τ), is
defined as:
Rxy (τ ) = lim
T →∞
1
2Tx(t)y(t − τ )dt
−T
T
∫
Telecommunication Systems Fundamentals
![Page 94: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/94.jpg)
Cross-Correlation from a Statistical Viewpoint
• Cross-Correlation of two stochastic processes can be
defined also using their joint-pdf:
RXY (t1, t2) = E X(t1)Y (t2)[ ]
= x−∞
∞
∫ yfX ( t1 ),Y (t2 )(x,y)dxdy−∞
∞
∫
• When both processes are stationary, then it holds that:
RXY (τ) = E X(t)Y (t − τ )[ ]
= x−∞
∞
∫ yfX ( t ),Y ( t−τ )(x, y)dxdy−∞
∞
∫
Telecommunication Systems Fundamentals
![Page 95: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/95.jpg)
Cross-Correlation Properties
– Rxy(τ) = Ryx(-τ).
– Pxy = Rxy(0) = Ryx(0) can be understood as the cross-power between
x(t) and y(t).
– The maximum of Rxy(τ) points to the time delay at which both
signals exhibit their maximum likeness
Telecommunication Systems Fundamentals
![Page 96: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/96.jpg)
Uncorrelated Processes
• Two SP are uncorrelated if Rxy(τ) = 0 for every value of τ
• It can be proven that if two SP are independent and at least one of then has zero mean, then they are uncorrelated
0
2
4
6
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1 - 1
0
1
Telecommunication Systems Fundamentals
![Page 97: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/97.jpg)
Sum of Signals
• Autocorrelation of the sum of two signal can be expressed as:
Rx + y (τ ) = R
x(τ ) + Ry (τ ) + R
xy (τ ) + Ryx (τ )
And particularizing for τ=0, the power of the sum is:
Px + y = R
x + y (0)
= R (0) + R (0) + R (0) + R (0)= Rx(0) + Ry (0) + R
xy (0) + Ryx (0)
= Px
+ Py + Pxy + Pyx
Important to note that only when the two processes are
uncorrelated, the power of the sum is the sum of the
powers
Telecommunication Systems Fundamentals
![Page 98: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/98.jpg)
Sine Signal plus DC
- 3 - 2 - 1 0 1 2 3- 3
- 2
- 1
0
1
2
3
Psin = limT →∞
1
2TA 2 sin 2 ( ft )dt
−T
T
∫
= limT →∞
1
2T−
1
2A
2 cos(2 ft) −1
2A
2 cos( ft − ft )
dt
−T
T
∫
= limT →∞
1
2T2T
1
2A
2=
A2
2
- 1
0
1
2
3
Pcon = limT → ∞
1
2TB
2dt
−T
T
∫ = B2
- 3 - 2 - 1 0 1 2 3- 3
- 2
conT → ∞ 2T −T
∫
- 3 - 2 - 1 0 1 2 3- 3
- 2
- 1
0
1
2
3
Pcon + sin = limT →∞
1
2T(B + A sin( ft)) 2
dt−T
T
∫
=A
2
2+ B
2
= Pcon + Psin
Telecommunication Systems Fundamentals
![Page 99: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/99.jpg)
Gaussian Noise
- 3
- 2
- 1
0
1
2
3
- 4
- 3
- 2
- 1
0
1
2
3
4
Zero mean B mena
- 3 - 2 - 1 0 1 2 3- 3
Pnoise = x2fX (x)dx
−∞
∞
∫= σ 2
- 3 - 2 - 1 0 1 2 3- 4
Pnoise+B = x2fX (x)dx
−∞
∞
∫= σ 2
+ B2
= Pnoise + Pcon
Telecommunication Systems Fundamentals
![Page 100: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/100.jpg)
Why do we want to study the Spectrum?
• Cyclic processes are quite common in nature
– Astronomy: lunar phases, planets orbits, solar storms, …
– Biology: heart beat,…
– Physics: acoustic vibrations, electromagnetic waves, …
• In communications, spectrum of the signal is of capital
importance when designing and analyzing systemsimportance when designing and analyzing systems
• If a signal is deterministic, the Fourier Transform
computes the spectrum, but what happen for
Stochastic Processes?
Telecommunication Systems Fundamentals
![Page 101: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/101.jpg)
Energy Spectral Density
• The Energy Spectral Density (ESD) of a energy-
defined signal is calculated as the Fourier Transform of
its auto-correlation function:
GX ( f ) = F RX (τ ){ }
= F x(τ ) ∗ x(−τ ){ }
– For Discrete time signals ESD is defined in similar way
= F x(τ ) ∗ x(−τ ){ }
= X ( f )2
Telecommunication Systems Fundamentals
![Page 102: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/102.jpg)
Energy Spectral Density
• The Energy Spectral Density describes how the energy
is distributed along different frequencies
• So, the total signal energy can be calculated by
integrating the ESD for all frequencies
E = x(t)2dt
∞
∫E X = x(t)2dt
−∞
∞
∫
= X ( f )2df
−∞
∞
∫
= GX ( f )df−∞
∞
∫
Telecommunication Systems Fundamentals
![Page 103: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/103.jpg)
Power Spectral Density
• The Power Spectral Density (PSD) of a power-defined signal is calculated as the Fourier Transform of its auto-correlation function:
SX ( f ) = F RX (τ ){ }
• For Discrete time signals PSD is defined in similar way
Telecommunication Systems Fundamentals
![Page 104: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/104.jpg)
Power Spectral Density
• The Power Spectral Density describes how the power
is distributed along different frequencies
• So, the total signal power can be calculated by
integrating the PSD for all frequencies
P = lim1
x(t)2dt
T
∫PX = limT →∞
1
2Tx(t)
2dt
−T
T
∫
= SX ( f )df−∞
∞
∫
Telecommunication Systems Fundamentals
![Page 105: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/105.jpg)
Sine Wave and the Sum of two Sine Waves
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1 - 1
0
1
0 . 5
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 1 . 5
- 1
- 0 . 5
0
0 . 5
1
1 . 5
0 . 5
1
- 5 - 4 - 3 - 2 - 1 0 1 2 3 4 5 - 0 . 5
0
- 4 0 0 0 - 3 0 0 0 - 2 0 0 0 - 1 0 0 0 0 1 0 0 0 2 0 0 0 3 0 0 0 4 0 0 00
1
2
3
4
5
6
7x 1 0
4
f r e q u e n c y ( H z )
- 5 0 5
x 1 0- 3
- 0 . 5
0
- 4 0 0 0 - 3 0 0 0 - 2 0 0 0 - 1 0 0 0 0 1 0 0 0 2 0 0 0 3 0 0 0 4 0 0 00
1
2
3
4
5
6
7x 1 0
4
f r e q u e n c y ( H z )
Telecommunication Systems Fundamentals
![Page 106: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/106.jpg)
6
7
8
9
0 0 . 0 0 2 0 . 0 0 4 0 . 0 0 6 0 . 0 0 8 0 . 0 1
- 8
- 6
- 4
- 2
0
2
4
6
Sine Wave and White Noise
- 5 0 5
x 1 0- 3
0
1
2
3
4
5
6
- 4 0 0 0 - 3 0 0 0 - 2 0 0 0 - 1 0 0 0 0 1 0 0 0 2 0 0 0 3 0 0 0 4 0 0 00
1
2
3
4
5
6
7x 1 0
4
f r e q u e n c y ( H z )
Telecommunication Systems Fundamentals
![Page 107: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/107.jpg)
PSD properties
1. Symmetry: PSDs are even functions, Gx(f)=Gx(-f) y
Sx(f)=Sx(-f).
2. Sign: PSD is real-valued and No-Negative for any
value of frequency f, Gx(f)≥0, Sx(f)≥0.
3. Integrability: Energy or Power can be calculated as
the integral of the their Spectral Density
Telecommunication Systems Fundamentals
![Page 108: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/108.jpg)
Cross Spectral Density
• Let be x(t) and y(t) two energy-defined signals which
cross-correlation function is Rxy(τ). Their Cross
Spectral Density is defined as the Fourier Transform of
their cross-correlation function:
Gxy ( f ) = F Rxy (τ){ }
• Analogously, if x(t) and y(t) are power-defined signals
their cross spectral density is defined as:
Gxy ( f ) = F Rxy (τ){ }
Sxy ( f ) = F Rxy (τ ){ }
Telecommunication Systems Fundamentals
![Page 109: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/109.jpg)
Sum of Signals
• Spectral Density of a sum of signals satisfies the
following relationship:
Sx + y ( f ) = F R
x + y (τ ){ }
{ }= F Rx(τ ) + Ry (τ ) + R
xy (τ ) + Ryx (τ ){ }
= Sx(τ ) + Sy (τ ) + S
xy (τ ) + Syx (τ )
Telecommunication Systems Fundamentals
![Page 110: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/110.jpg)
Summary of Spectral Density
• Energy (or Power) Spectral Density describes how energy (or power) is distributed along frequencies
• Can be used for Stochastic Processes and deterministic signals
• In the case of SP, Spectral Density represents an statistical average of the process. One particular realization may have different spectrum
• Spectral Density of the sum of several signals can be calculated by computing the Spectral Density of each elementary signal and their cross spectral density
Telecommunication Systems Fundamentals
![Page 111: Topic1 Review of Stochastic Processes · • Review of Statistical Basics: Modeling of Stochastic Processes – Amplitude distribution (probability density function, pdf) and averages](https://reader036.fdocuments.us/reader036/viewer/2022081612/5f3f56e1a764c908b85eb8cb/html5/thumbnails/111.jpg)
Summary of Concepts in this Chapter
• In this chapter we learned:
– How to model a Stochastic Process
– Different properties of SP: Independence, Stationarity, Ergodicity
– How to compute the autocorrelation of a SP and its physical meaning
– How to compute the Energy(Power) Spectral Density of a SP and its
physical meaningphysical meaning
Telecommunication Systems Fundamentals
111