Beruflich Dokumente
Kultur Dokumente
Two Highly
Negatively
Correlated
DT Signals
Thermometer
Correlograms Correlograms
Signals Signals
Correlogram Correlogram
Two
Two Partially
Uncorrelated
Correlated
CT Signals
DT Signals
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 5 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 6
1
Correlograms Correlograms
Signals
DT Sinusoids With CT Sinusoids With
Correlogram a Time Delay a Time Delay
Signals Signals
These two CT
signals are not Correlogram Correlogram
strongly
correlated but
would be if one
were shifted in
time the right
amount
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 7 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 8
Two Non-
Linearly Related
DT Signals
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 9 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 10
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 11 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 12
2
The Correlation Function Correlation of Energy Signals
Positively Uncorrelated Random Negatively
The correlation between two energy signals, x and y, is
Correlated Random DT Signals with Correlated Random
the area under (for CT signals) or the sum of (for DT
DT Signals with Non-zero Mean DT Signals
signals) the product of x and y.
Non-zero Mean with Non-zero Mean
" "
# x(t ) y (t )dt
*
or # x[n]y [n] *
!" n=!"
"# n=!"
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 13 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 14
function. If the two signals are both periodic and their fundamental
R xy (! ) = x( "! ) # y(! ) or R xy [ m ] = x[!m] " y[ m ] periods have a finite least common period,
1 1
N n=!N
R xy (! ) = x(t ) y(t + ! )dt or R xy [ m ] = x[ n ] y[n + m ]
T "T
It also follows that
R xy (! ) "$# X ( f ) Y( f )
F *
or R xy [ m ]!#" X ( F ) Y( F )
F *
where T or N is any integer multiple of that least common
period.
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 15 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 16
R xy (! ) "$
FS
# X*[ k ] Y[ k ] or R xy [ m ]!#
FS
" X* [ k ] Y[ k ]
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 17 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 18
3
Autocorrelation
Correlation of Sinusoids A very important special case of correlation is
autocorrelation. Autocorrelation is the correlation of a
function with a shifted version of itself. For energy signals,
# "
!" n=!"
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 23 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 24
4
Matched Filters Matched Filters
It has been shown
that the optimal
• A very useful technique for detecting the filter to detect a
presence of a signal of a certain shape in the noisy signal is one
whose impulse
presence of noise is the matched filter. response is
• The matched filter uses correlation to detect proportional to the
time inverse of the
the signal so this filter is sometimes called a signal. Here are
correlation filter some examples of
waveshapes
• It is often used to detect 1’s and 0’s in a encoding 1’s and
binary data stream 0’s and the impulse
responses of
matched filters.
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 25 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 26
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 27 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 28
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 29 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 30
5
Autocorrelation Examples Autocorrelation Examples
• One way to simulate a random signal is with
a summation of sinusoids of different Let a random signal be described by
frequencies and random phases N
x( t ) = # Ak cos( 2!f0k t + " k )
• Since all the sinusoids have different k =1
frequencies the autocorrelation of the sum is Since all the sinusoids are at different frequencies,
simply the sum of the autocorrelations N
R x (! ) = " R k (! )
• Also, since a time shift (phase shift) does not k =1
affect the autocorrelation, when the phases where R k (! ) is the autocorrelation of Ak cos( 2!f0 k t + " k ) .
are randomized the signals change, but not
their autocorrelations
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 31 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 32
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 33 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 34
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 35 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 36
6
Cross Correlation Cross Correlation
Cross correlation is really just “correlation” in the cases in
which the two signals being compared are different. The A comparison of x and y with y shifted for maximum
name is commonly used to distinguish it from autocorrelation. correlation.
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 37 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 38
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 39 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 40
# x( t ) # X( f ) # x[n] = $ X(F ) system and the ESD of the response of the system. It can be
2 2
Ex =
2 2
Ex = dt = df or dF
!" !" n=!"
1 shown (pp. 606-607) that they are related by
!y ( f ) = H( f ) !x ( f ) = H( f ) H* ( f )!x ( f )
2
The quantity, X( f ) , or X( F ) , is called the energy spectral
2 2
!x ( f ) = X( f ) !x ( F ) = X( F )
2 2
or
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 41 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 42
7
Energy Spectral Density Energy Spectral Density
R x ( t ) !#
F
" $x ( f ) R x [ n ]!# " $x ( F )
F
or
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 43 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 44
or
G y ( F ) = H( F ) G x ( F ) = H( F ) H* ( F ) G x ( F )
2
5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 45 5/2/05 M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 46
Typical
Signals in
PSD
Concept