A novel adaptive noise canceller with master-slave structure

This paper presents an efficient adaptive noise canceller (ANC) with master-slave structure. The proposed ANC has two adaptive filters: a master filter (MF) and a slave filter (SF). The SF is used to estimate the signal-to-noise ratio (SNR) of input signals, which controls the step-size of the MF. With regard to this controlling function, we have experimented on five different types of elementary functions. Computer simulation results demonstrate that this proposed ANC can achieve the same good performance as the previous MSANC, but has a smaller computation. In addition, the positive exponential function is the optimal controlling function for the step-size of the MF