Conf. on Info. Sciences and Systems, Baltimore, MD, March 2001
Exploiting Sparsity in Adaptive Filters
R.K. Martin, W.A. Sethares, R.C. Williamson and C.R. Johnson, Jr.
Abstract
This paper studies a class of algorithms called Natural Gradient (NG)
algorithms and their approximations, known as ANG algorithms. The LMS
algorithm is derived within the NG framework, and a family of LMS
variants that exploit sparsity is derived. Mean squared error (MSE)
analysis of the family of ANG algorithms is provided, and it is shown
that if the system is sparse, then the new algorithms will often
converge faster for a given total asymptotic MSE. Simulations are
provided to confirm the analysis. In addition, Bayesian priors
matching the statistics of a database of real channels are given.
Actual channels are identified by algorithms that exploit these priors
and by LMS, showing a realistic application of these algorithms.
