 Designed leastsquares
algorithm that combines, in an optimal manner, data arising
from a finite collection of uncertain models. The algorithm
can take into account data uncertainties with different
sophistication levels. The algorithm demonstrated improved
performance when it was applied to fusion of data arriving
from a distributed network of sensors with varying degrees
of reliability. The Algorithm was also applied to diversity
combining of signals in the presence of microscopic or macroscopic
fading.
 Developed adaptive
algorithm with optimum error nonlinearity in the adaptation
equation. Nonlinearity is a function of the pdf of the additive
noise. Algorithm attains a lower steadystate error compared
with adaptive algorithms employing other nonlinearities.
Research resulted in 4 conference publications.



