Evaluation of multiple reference active noise control algorithms on Dornier 328 aircraft data

摘要

This article presents an evaluation of multiple reference adaptive algorithms. Two least mean squares-types (LMS-types) and a Newton-type algorithm are considered. The special structure of the adaptive filtering problem implies that the Newton-type algorithm can be implemented with the same numerical complexity as LMS-type algorithms. The concept of a first filtered-x Newton algorithm is thus introduced.