DocumentCode :
1409661
Title :
Robust relative stability of time-invariant and time-varying lattice filters
Author :
Dasgupta, Soura ; Fu, Minyue ; Schwarz, Chris
Author_Institution :
Dept. of Electr. & Comput. Eng., Iowa Univ., Iowa City, IA, USA
Volume :
46
Issue :
8
fYear :
1998
fDate :
8/1/1998 12:00:00 AM
Firstpage :
2088
Lastpage :
2100
Abstract :
We consider the relative stability of time-invariant and time-varying unnormalized lattice filters. First, we consider a set of lattice filters whose reflection parameters αi obey |αi|⩽δi and provide necessary and sufficient conditions on the δi that guarantee that each time-invariant lattice in the set has poles inside a circle of prescribed radius 1/ρ<1, i.e., is relatively stable with degree of stability ln ρ. We also show that the relative stability of the whole family is equivalent to the relative stability of a single filter obtained by fixing each αi to δi and can be checked with only the real poles of this filter. Counterexamples are given to show that a number of properties that hold for stability of LTI Lattices do not apply to relative stability verification. Second, we give a diagonal Lyapunov matrix that is useful in checking the above pole condition. Finally, we consider the time-varying problem where the reflection coefficients vary in a region where the frozen transfer functions have poles with magnitude less than 1/ρ and provide bounds on their rate of variations that ensure that the zero input state solution of the time-varying lattice decays exponentially at a rate faster than 1/ρ1>1/ρ
Keywords :
Lyapunov matrix equations; circuit stability; filtering theory; lattice filters; poles and zeros; time-varying filters; transfer functions; circle; diagonal Lyapunov matrix; exponential decay; necessary conditions; radius; real poles; reflection coefficients; reflection parameters; robust relative stability; sufficient conditions; time-invariant lattice filter; time-varying lattice filter; transfer functions; unnormalized lattice filters; zero input state solution; Delay; Lattices; Linear predictive coding; Nonlinear filters; Poles and zeros; Reflection; Robust stability; Speech processing; Sufficient conditions; Transfer functions;
fLanguage :
English
Journal_Title :
Signal Processing, IEEE Transactions on
Publisher :
ieee
ISSN :
1053-587X
Type :
jour
DOI :
10.1109/78.705417
Filename :
705417
Link To Document :
بازگشت