A Tuning-Less Approach in Secondary Path Modeling in Active Noise Control Systems

This correspondence proposes a new approach to the problem of secondary path estimation in active noise control (ANC) systems aimed to avoid any tuning during system setup. To meet this target, adaptation rules are proposed, which do not have any parameters to be tuned and whose behavior can be predicted with great accuracy, because analytical expressions of their modeling mean square relative error curves are given. Computer simulations show that their performances are comparable with those obtained by using optimal variable step size (VSS) versions of the least mean square (LMS) algorithm.