Weak Convergence and Asymptotic Properties of Adaptive Filters with Constant Gains.
Abstract
This paper illustrates the power of weak convergence methods through the analysis of the basic algorithm of adaptive filtering. Except for the simplest cases (e.g., when (y sub n, psi sub n) are mutually independent), the analysis of the algorithm for fixed epsilon is difficult. However, asymptotic analysis (epsilon approaches limit of O) via weak convergence methods provides much information, relatively painlessly.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 06, 1983
- Accession Number
- ADA129214
Entities
People
- Adam Shwartz
- Harold J. Kushner
Organizations
- Brown University