A new adaptive control strategy for hybrid narrowband active noise control systems in a multi-noise environment

Abstract

In this paper, a new hybrid narrowband active noise control (HNANC) system which effectively counters noise from multiple sources is proposed. Usually, active noise control (ANC) systems fail in the event of encountering uncorrelated noise, since such systems mostly are designed to operate when disturbances correlated to the reference signal are present. To deal with the multiple-noise scenario in an industrial environment, a new adaptive control strategy is introduced consisting of three subsystems: 1) a narrowband feedforward controller where we use filtered-x weighted accumulated LMS (FXWALMS) in combination with a new variable momentum factor scheme, 2) a feedback ANC system adapted by the FXLMS algorithm to specifically counter uncorrelated disturbances, and 3) the supporting adaptive filter that has a twofold objective, firstly, to provide appropriate error/reference signals to the noise cancelling adaptive filters and secondly, to provide the most accurate energy estimate of the constrained error signal for the variable momentum factor to improve the steady state performance of the feedforward controller. For improved adaptive filter convergence, we also reduce the number of reference signal filtering blocks in the feedforward NANC controller. A detailed mathematical analysis of the hybrid structure is performed to establish the basis for the proposed algorithm. Extensive simulations show that the proposed HNANC system gives an improved overall ANC performance over that of existing counterparts.