Scribd
Upload
340 views15 pages

Statistical Signal Processing

Statistical signal processing (SSP) deals with signals that have inherent uncertainties and unknowns. It uses probability and statistics to model these uncertainties and analyze the signals. The key areas of SSP include estimation, filtering, and detection. SSP is applied to real-world signals like seismic signals and radar signals. It observes data through experiments and uses statistical tools like probability distributions, random variables, and measures of central tendency and variation to characterize the data and interpret the results.

Uploaded by

suren
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
340 views15 pages

Statistical Signal Processing

Statistical signal processing (SSP) deals with signals that have inherent uncertainties and unknowns. It uses probability and statistics to model these uncertainties and analyze the signals. The key areas of SSP include estimation, filtering, and detection. SSP is applied to real-world signals like seismic signals and radar signals. It observes data through experiments and uses statistical tools like probability distributions, random variables, and measures of central tendency and variation to characterize the data and interpret the results.

Uploaded by

suren
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 15

Statistical Signal Processing

DSP Vs SSP

Deterministic Signal Processing Statistical Signal Processing

Assumes complete knowledge Believes that signal has


about the signal uncertainties and unknowns
True in real world
signals like (i) Seismic
Probability
signal (ii) Radarand
signalstatistics are used
Processes digital signals using etcaddress uncertainties and
algorithms to
unknowns
To capture the
Probability uncertainties in
model the data/signal
By conducting
Collect data experiment.
Statistics
Statistics
To interpret
is a
cannot depend
function
the probability
of
Statistics on unknown
observed
model data.
quantities.
Areas of SSP

Vital Areas of SSP

Estimation Filtering Detection

Real time estimation of


Surface
Observe of
reflected streaming data
Send acoustic
earth
wave
wave (Tracking: Kalman filter)
Estimate the distance

Oil pool
Review of fundamental concepts
Outcome cannot be predicted
with certainty
Random Result
Experiment Outcome
1st order 2nd order Higher order
moment moment moment
Mapping Sample Set of all possible
space to a point outcomes
on real line
Random variable Sample space
Types of Random variable
Subset of sample space
Continuous Discrete Mixed Event Sure Event
RV RV RV
Characterized Impossible
by Types of
Event
Probability Probability Events
Independent
Density function Mass function
Event

Mutually
Exclusive Event
Probability
• Tool to handle the uncertainty.
• Quantify the uncertainty in the range [0,1].
Terminology in Set theory and Probability
theory

S.I.No Set theory terminology Probability theory


terminology
1 Universe Sample space
2 Element Outcome
3 Set Event
Probability Distribution function
• Notation: FX (x)
• Definition (CDF): FX ( x)  Pr( X  x)

• PMF (Discrete Random variable): p X (x)


• Types of distribution :

(i) Continuous and (ii) Discrete

• Examples of continuous distributions


Uniform, Normal, Exponential etc

• Examples of Discrete distributions


– Bernoulli, Bionomial, Poisson, Hypergeometric etc.
Example of CDF

• Random Experiment: Tossing a fair coin thrice


• Sample Space:
{(TTT)(TTH)(THT)(THH)(HTT)(HTH)(HHT)(HHH)}
• Random Variable (X): Number of Tails
X={0,1,23}
1 3 3 1 
• PMF: p
X (x )   , , , 
8 8 8 8 
• CDF : p
X (x ) 
1 4 7 8 
 , , , 
8 8 8 8 
Plot of CDF
FX (x)
8
8

7
8

4
(i) CDF is right
8 continuous function
1 (ii) CDF is a non-
8
decreasing function

0 3 x
1 2
Measure of data properties

Data properties

Central Variation Shape


Tendency Skew
Range
Mean Standard
Median deviation
Mode Variance
Data properties (Cont.,)

Central Tendency
(Location)

Variation (Dispersion)

Shape
Statistical Mean (or) Expectation
• For discrete random variable ‘X’, the mean
value is given by
E ( X )   xi p X ( xi )
i

• For continuous random variable ‘X’ the mean


value is given by

E( X )   xf

X ( x)dx
Properties of Expectation operator
• Expectation is a linear operator.

E (aX )  aE ( X )
E ( X  Y )  E ( X )  E (Y )
E (c )  c
Variance

Var ( X )  E ( X )  E ( X )
2 2

Var(aX )  a Var( X )
2

Var (c )  0

Var ( X  Y )  Var ( X )  Var (Y )


Random process/Stochastic process
• Collection of “indexed” random variable.
• Index could be time, space etc.

You might also like