Higher Order Statistics for Estimation of Degree of Noncausality in Active Noise Control System

Abstract

The paper highlights the effects of degree of noncausality in the performance of single channel Active Noise Control (ANC) system. It is shown that for the optimal positioning of the secondary source (loudspeaker) the degree of noncausality calculation is important. It is also restated that the empirical formula found by Elliot et al [1,2] for the maximum convergence coefficient, which is used in the filtered-x LMS algorithm, is important to be evaluated. Hence to evaluate the degree of noncausality and to find out the Elliot's approximate convergence coefficient, the primary and the secondary path delays are required to be estimated apriori. A novel higher order statistics (HOS) based method is proposed which is efficient to estimate the time delay or sample delay of the acoustic primary path as well as the electro-acoustic secondary path even in the presence of measurement noise with small data length.