Recursive parameter estimation methods: An adaptive filtering perspective

Abstract

Listen

Translate article

This tutorial provides an overview of recursive parameter estimation methods from the viewpoint of the adaptive filtering field. Specific topics to be discussed include gradient versus least-squares techniques, adaptive infinite-impulse response filters, and bias removal methods. Particular emphasis is placed on the analytical behaviors of the algorithms, their estimation characteristics, and their tracking abilities in mildly nonstationary environments. Applications of the methods to identification problems in acoustics are discussed. [This work has been supported by NSF Grant No. MIP-9501680.]