Online secondary path modeling algorithm without auxiliary noise for narrowband active noise control

Abstract

Online secondary path modeling (SPM) is a practical method for real-time noise reduction in narrowband active noise control (NANC) systems, particularly when addressing variations in the secondary path. However, the common practice of using auxiliary noise for online SPM increases the residual noise power and deteriorates the noise reduction performance. The present study proposes a strategy that does not rely on auxiliary noise for online SPM in NANC systems. The proposed algorithm comprises two stages: Stage A models the primary path, whereas Stage B concurrently engages in online SPM and active noise control. The control signal is used to model the discrete Fourier transform (DFT) coefficients of the secondary path, avoiding the need for an auxiliary noise and significantly reducing the computational complexity. Moreover, the predicted primary path from Stage A is employed to obtain the pure desired signal of the online SPM. This strategy decorrelates the primary noise and the modeling signal, and accelerates the convergence of the algorithm. Simulations of recorded data demonstrate that the proposed algorithm can quickly track variations in both the primary and secondary paths, and maintain the noise reduction performance and stability of the system.