Efficient combination of feedforward and feedback structures for nonlinear narrowband active noise control

•Analyzing the primary noise of NNANC system with nonlinear primary path and non-minimum phase secondary path, it is found that the frequency components of primary noise can be predicted by reference signal.•Proposing a novel NNANC scheme using the combination of feedforward and feedback structures.•Filtered-error structure is added to reduce the computational complexity.•Improved scheme using error separation technique is developed to outperform on convergence speed and steady-state error.

In nonlinear narrowband active noise control (NNANC) system, there often exist nonlinear distortions in such cases: the primary path is nonlinear and the secondary path is non-minimum phase. In order to solve these nonlinear issues, a novel NNANC scheme based the combination of feedforward and feedback structures is proposed in this paper. In new NNANC scheme, the spectrum information fed back from the noise cancelling point is corrected by the frequency components expanded from the reference signal to obtain precise frequency components of primary noise. Then these frequency components are used as the input signal of new NNANC scheme and each frequency component is processed independently to reduce the undesired noise. For reducing the computational complexity and simplifying the implementation structure of new NNANC scheme, the filtered-error structure is used in proposed scheme to form Proposed-1 algorithm. To improve the performance of Proposed-1 algorithm even further, error separation technique using bandpass filter banks for these frequency components is developed to form Proposed-2 algorithm. Computational complexity analysis and numerous simulations for different narrowband noise and nonlinear primary path are given to demonstrate the effectiveness of proposed schemes.