Scribd
Upload
63 views4 pages

XN Xnim I K N M: The Mean Value Is

The document summarizes Bartlett's method for reducing the variance in the periodogram of a signal. It involves (1) dividing the signal into non-overlapping segments, (2) computing the periodogram of each segment, and (3) averaging the periodograms to obtain the Bartlett power spectrum estimate. The variance of the Bartlett estimate is reduced by a factor of K, where K is the number of segments.

Uploaded by

jayeshdiwan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
63 views4 pages

XN Xnim I K N M: The Mean Value Is

The document summarizes Bartlett's method for reducing the variance in the periodogram of a signal. It involves (1) dividing the signal into non-overlapping segments, (2) computing the periodogram of each segment, and (3) averaging the periodograms to obtain the Bartlett power spectrum estimate. The variance of the Bartlett estimate is reduced by a factor of K, where K is the number of segments.

Uploaded by

jayeshdiwan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

Bartletts method is used for reducing the variance in the periodogram.

It involves three
steps. (1) The N - point sequence is subdivided in to K nonoverlapping segments. Here each
segment has length M. This results in the K data segments.

xi [n] x[n iM ] i 0,1, ..., K 1


n 0,1, ..., M 1
The periodogram for each segment will be,
M 1

1
Pxx( i ) ( f )
M

x [n]e

2
j 2 fn

n 0

i = 0,1,...,K 1

We can compute the Bartlett power spectrum estimate by averaging periodogram computed for
each segment as,

1
P (f)
K
B
xx

K 1

(i )
xx

i 0

(f)

The mean value is,

1
E[ P ( f )]
K
B
xx

K 1

E[ P
i 0

(i )
xx

( f )]

E[ Pxx(i ) ( f )]

It shows that expected value of Bartlett periodogram is same as expected value of


periodogram for each segment.

E[ P ( f )]
(i )
xx

M 1

m ( M 1)

m
j 2 fm
xx ( m)e
M

12

sin ( f ) M

)
1 2 xx sin ( f ) d

1 sin fM
WB ( f )

M sin f

WB ( f ) 1
= 0,

m
M

, for m M 1
otherwise

Variance of Bartlett estimate can be computed as,

1 K 1
var[P ( f )] 2 var[Pxx(i ) ( f )]
K i 0
B
xx

1
var[Pxx( i ) ( f )]
K

sin 2 fM
1
var[Pxx( B ) ( f )] 2xx ( f ) 1

K
M
sin
2

The variance of Bartlett power spectrum estimate is reduces by factor K.

Welch method.

xi [n] x[n iD ]

n 0,1,..., M 1
i 0,1,..., L 1

1
(i )
P%
(
f
)

xx
MU

1
U
M
(W )
xx

M 1

M 1

x [n]w[n]e
n 0

2
j 2 fn

i = 0,1,...,L 1

w [ n]
2

n 0

1 L1 %
( f ) Pxx( i ) ( f )
L i 0

(W )
xx

E[ P

1 L1 %
( f )] E[ Pxx( i ) ( f )]
L i 0

B & T method

P (f)
BT
xx

M 1

m ( M 1)

E[ P ( f )]
BT
xx

rxx [m]w[m]e j 2 fm

M 1

m ( M 1)

E[rxx [m]]w[m]e j 2 fm

M 1

m ( M 1)

xx [m]wB [m]w[m]e j 2 fm

var[ PxxBT ( f )] E{[ PxxBT ( f ) 2 ]} {E[ PxxBT ( f )]}2

1
var[ P ( f )] ( f )
N
BT
xx

2
xx

2
w
[
m

]
m ( M 1)

M 1

You might also like