|
||||||||||||||||||||||||||||||
So far in the above discussion we have discussed about the convolution of two continous time signals which is used to find the output y(t) of a system. Correlation is a mathematical operation that closely resembles convolution. Correlation is basically used to compare two signals. Correlation is the measure of the degree to which two signals are similar. The correlation of two signals is divided into two ways: (i) Cross-correlation, (ii) Auto-correlation.
Cross correlation is a measure of similarity between two waveforms as a function of time gap or delay applied to one of them. The cross correlation between a pair of continuous time signals f(t) and g(t) is given by where --------- (1) and it can be derived for discrete time signal f(n) and g(n) as where The index and are the shift parameters for continuous time and discrete time signals respectively. The order of subscripts 'fg' indicates that f(t) or f(n) are the reference sequence in continuous-time and discrete-time respectively that remains unshifted in time whereas the sequence g(t) or g(n) are shifted '' or 'k' units in time with respect to f(t) or f(n) respectively. If we want to fix g(t) and to shift f(t), then the correlation of two sequences can be written as
-----------------------(2) and for discrete time signals, the cross-correlation can be written as
Now comparing eqn. (1) and eqn. (2) we find that
and similarly for discrete time signal
Where for continuous time signals is the folded version of about =0 and for discrete time signal is the folded version of about k=0. Now we can rewrite the eqn. (1) as
and similarly for discrete time signal
From the above equation(3) we find that the correlation process is essentially the convolution of two data sequence in which one of the sequence has been reversed.
Auto correlation of a continuous time signal is the correlation of the signal with itself. The auto correlation of a continuous time signal f(t) is defined as
for discrete time signal f(n) it is defined as
or equivalently we can write and similarly for discrete time signal f(n) we can write
Relation to Signal energy and Signal power:
The auto-correlation function of a periodic signal is itself a periodic signal with a period the same as that of the original signal. If f(t) is an energy signal, its auto-correlation is
and for continuous-time and discrete-time signals respectively. After applying a zero shift it becomes or which is the total signal energy of the signal. If f(t) or f[n] is a power signal, the auto-correlation at zero shift is
for continuous-time and discrete-time respectively which is the average signal power of the signal.
Properties of Auto-correlation:
The auto-correlation depends on the choice of the amount of shift applied. we can say from the observation that the value of the auto-correlation can never be bigger than it is at zero shift. That is,
It will happen because at a zero shift, the correlation with itself is obviously as large as it can get since the shifted and unshifted versions coincide. Another property of auto-correlation function is that all auto-correlation functions are even functions (but not all correlation functions). or If we make the change of variable and We can show that It can be shown by a similar technique for discrete-time signal also
Another characteristic of auto-correlation function is that a time shift of a signal does not make any change of its auto-correlation. Letbe the auto-correlation function of a discrete-time energy signal f[n]. Then
Now let y[n] = f [n-n0] . Then
Now we can make a change of variable q = n-n0 . Then
Another characteristic of auto-correlation function is that the auto-correlation of a sum of sinusoids of different frequencies is the sum of auto-correlation of the individual sinusoids. To demonstrate this idea let a continuous-time power signal f(t) be a sum of two sinusoids f1(t) and f2(t), where
The auto-correlation of this signal is
Therefore,
Cite this Simulator: ..... |
..... |
..... |
|
|||||||||||||||||||||||||||
Powered by AmritaVirtual Lab Collaborative Platform [ Ver 00.13. ]
|