A Simple on-Line Period Estimation for an Iterative Learning Active Noise Control Scheme

The paper describes a period/frequency adaptation mechanism applied to an active noise control scheme designed to attenuate disturbances with high autocorrelation characteristics. The proposed modification tracks slowly varying changes in the noise autocorrelation function peak. Moreover, it can switch between different algorithm settings when an abrupt change in noise characteristics is detected. This modification increases the method’s robustness and the area of potential implementations. The algorithm’s behavior and performance are verified with computer simulations using real-world signals and acoustic path models identified experimentally. The results confirm that the previously proposed ANC algorithm can be extended with the period tracking mechanism using little additional computational resources and without apparent degradation in attenuation or stability.