A new fully adaptive algorithm with assured stability for adaptive active noise control

Abstract
A stability-assured adaptive algorithm of updating the feedforward controller for a general adaptive active noise control system is investigated based on the strictly positive real (SPR) property of a error model. In particular, when there is uncertainly changeable channel dynamics between the control speaker and the objective area and there exists a feedback path from the control speaker to the detection microphone, almost all conventional adaptive algorithms fail to work. Two type of fully adaptive algorithms are proposed which can work in the condition that all the channel dynamics are unknown and changeable, without using any additional microphones and loudspeakers, and without identification of additional models or with assured stability.

This publication has 4 references indexed in Scilit: