This preview shows pages 1–16. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Chapter 11 Spectral Representation Teacher: Office: 805 Tel: ext. 31822 Email: schen@mail.nctu.edu.tw 2 11.1 Factorization and Innovations x & real WSS RP ( ) t X minimumphase system ( ) L s L output with input e white noise process ( ) t i i signal modeling i h ( ) t X i i regulare Minimumphase system ( ) L s L causal e impulse response ( ) t finite energy (i.e., causal and stable)e inverse system 1 ( ) ( ) s L s = causal e ( ) t finite energy 3 ( ) L s is minimumphase ( ) L s and ( ) s are analytic in the right splane, i.e., all poles of ( ) L s and ( ) s are in the left splane. 4 A RP e regular if it is linearly equivalent with a whitenoise process ( ) t i in the sense ( ) ( ) ( ) t t d = i X e ( ) ( ) R = ii ( ) ( ) ( ) t t d = X i { } 2 2 ( ) ( ) E t t dt = < X ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) R R = = = XX ii ( ) ( ) d = + e ( ) ( ) R R = XX XX 5 ( ) ( ) ( ) S s L s L s = 2 ( )  ( ) S L = ( ) L s L innovations filter of ( ) t X i 1 ( ) ( ) s L s = whitening filter of ( ) t X i ( ) t i i innovations of ( ) t X 6 h& ( ) L s bA Given a positive even function ( ) S a minimumphase system ( ) L s s.t. 2  ( )  ( ) L j S = h PaleyWiener Conditione 2 log ( ) 1 S d  < + e ( ) S ( ) S line e BL( predictable) 7 h& in general for special casee e ( ) L s L rational function of sL ( ) S 2 rational functione e 2 * ( )  ( ) ( ) ( ) ( ) ( ) S L j L j L j L j L j = = =  ( ) ( ) S S =  e 2 2 ( ) ( ) ( ) A S B = 2 2 ( ) ( ) ( ) A s S s B s = 8 j s s = j s (e * j s L * j s for complex j s ) e ( ) S s L poles e zerose Fig. 112(a)e e e poles e zeros L ( ) ( ) ( ) ( ) ( ) N s N s S s D s D s = ( ) ( ) L s L s = e ( ) ( ) N s D s L poles e zerose e Fig. 112(b)e ( ) ( )/ ( ) L s N s D s = 9 10 Ex. 111 2 2 ( ) N S = + ( ) L s Solutione 2 2 ( ) ( )( ) N N S s s s s = = + ( ) N L s s = + 11 Ex. 112 2 4 2 49 25 ( ) 10 9 S + = + + 2 4 2 2 2 49 25 (7 5 )(7 5 ) (7 5 )(7 5 ) ( ) 10 9 (1 )(9 ) (1 )(1 )(3 )(3 ) s s s s s S s s s s s s s s s + + = = = + + + 7 5 ( ) (1 )(3 ) s L s s s + = + + 12 Ex. 113 4 25 ( ) ( 1) S = + 4 2 2 25 25 ( ) ( 1) ( 2 1)( 2 1) S s s s s s s = = + + + + 2 5 ( ) ( 2 1) L s s s = + + 13 Discretetime Processes A real WSS discretetime process [ ] n X is regular if ( ) S z can be written as 1 ( ) ( ) ( ) S z L z L z = 2 ( )  ( )  j j S e L e = where ( ) L z is a minimumphase systeme 14 [ ] [ ] [ ] k n k n k = =  i X i [ ] [ ] R m m = ii [ ] [ ] [ ] k n k n k = =  X i [ ] { } [ ] 2 2 k E n k = = < X 15 h& [ ] n X is linearly equivalent with a whitenoise process...
View Full
Document
 Fall '08
 SinHorngChen

Click to edit the document details