Orthogonal adaptation for active noise control

Abstract

Many active noise control (ANC) systems apply the filtered-x least mean squares (FXLMS) algorithm for controller adaptation. The accuracy of path models is an important issue in these systems. Since parameter drifting in a noise field may cause model error between the secondary path and its prestored model in an ANC system, some ANC systems employ two adaptive processes for path modeling and controller adaptation respectively. In this paper, a new ANC system is proposed with adaptive path modeling and nonadaptive controller design. The proposed ANC system is noninvasive without persistent excitations. It avoids the slow convergence and inevitable estimation errors in controller adaptation. A rigorous analysis is presented to prove that the new ANC system will converge to an optimal one in the minimum H2 norm sense. Experimental results are presented to verify the performance of the proposed ANC system.