| | |
Summary: MEAN-SQUARE ANALYSIS OF NORMALIZED LEAKY ADAPTIVE FILTERS
ALIH. SAYED' AND TAREQY. AL-NAFFOURI~
'Electrical Engineering Department
University of California, Los Angeles, CA 90095
ElectricaI Engineering Department
Stanford University, CA 94305
ABSTRACT
In this paper, we study leaky adaptive algorithms that em-
ploy a general scalar or matrix data nonlinearity. We per-
form mean-squareanalysis of this class of algorithmswithout
imposing restrictions on the distribution of the input signal.
In particular,we deriveconditions on the step-size for stabil-
ity, and provide closed form expressions for the steady-state
performance.
1. ADAPTIVE FILTERING MODEL
In this paper, we consider the following class of leaky
adaptive filters:
Wi+l -- (1- cup)wi+pEt(ui)uTe(z) (1)
e(i) = d(i) - uiwi (2)
d(i) = uiwo+v(2) (3)
|