Improved Parallel Feedback Active noise control using linear prediction for adaptive noise decomposition

This paper presents an improved Parallel Feedback Active noise control (PFANC) method with adaptive noise signal decomposition using NLMS based linear prediction. The proposed method is tested for several stationary periodic signals, quasi-periodic signals and non-stationary signals buried in Additive White Gaussian noise (AWGN) under different Signal to Noise Ratios (SNR). Performance improvement in terms of Noise Attenuation Levels (NAL) exceeding 40dB have been obtained and compared with previous methods. Proposed method has lower implementation cost, lower computational delay, better noise tracking capability and better NAL performance as compared to previous methods.