Design of Decoupling Active Noise Control Systems H.Hamidi, F. Taringoo, A. Nasiri, Islamic Azad University of Boroujerd ABSTRACT: In this paper a simple method for multi- channel Active Noise Control is introduced. The proposed structure is a combination of an adaptive FIR filter, based on the FXLMS algorithm, and the ALIGN algorithm resulting in decoupling matrix for secondary path transfer function. This structure converts a multi-channel ANC systems to decoupled systems which results reduction in computational load without affecting the convergence behavior of the FXLMS system Key words Multi channel Active Noise Control, Adaptive control, ALIGN algorithm, Decoupling 1. Introduction In Active Noise Control (ANC) systems, acoustical noise is tried to be cancelled (in a desired point or zone) by the anti noise signal generated by controlled loudspeakers [1],[2]. In ANC applications a common approach is to use adaptive FIR filters trained with the Filtered X_ LMS algorithm [1]. The block diagram of an FXLMS is shown in figure (1). In this figure it is clear that the outputs of adaptive filters are filtered by the secondary path transfer function from loudspeakers to error microphones. In this paper we assume equality between number of loudspeakers and error microphone as a result of which the secondary path transfer function matrix is a square matrix. It is well-known that in ANC systems, multi- channel structures involve much more computational complexity compared to single- channel ones. This is due to having more than one loudspeaker and microphone in a multi-channel structure, and that each loudspeaker affects all microphones. Much effort has been made to reduce the computational complexity and to improve the convergence of the adaptive algorithm in multi- channel systems [3]-[6]. In [3], a mathematical technique based on the spectral factorization of the secondary path transfer function matrix has been used to improve the convergence of the FXLMS algorithm. In[4] a modified algorithm has been proposed to reduce the computational complexity , whereas in [5], applying an inverse matrix for the FIR filters has shown to do both tasks. In [6], it has been shown that if a condition is met for the secondary path transfer matrix, it would simplify the algorithm for computing the adaptive filter coefficients, while if not met, it would cause instability. Our method is based on applying the Align matrix introduced in control theory to the ANC systems [7] In the proposed method, a gain matrix is applied between adaptive filters outputs and loudspeakers. This gain matrix is obtained according to the ALIGN algorithm [8] applied to secondary transfer function in noise frequency in order to decouple the secondary path transfer matrix. 2. Description of the Multichannle ANC[1] The general multiple-reference/ multiple-output ANC system using the FXLMS algorithm is shown