Adjoint EKF learning in recurrent neural networks for nonlinear active noise control

Deep ANC: A deep learning approach to active noise control

Traditional active noise control (ANC) methods are based on adaptive signal processing with the least mean square algorithm as the foundation. They are linear systems and do not perform satisfactorily in the presence of nonlinear distortions. In this paper, we formulate ANC as a supervised learning problem and propose a deep learning approach, called deep ANC, to address the nonlinear ANC problem. The main idea is to employ deep learning to encode the optimal control parameters corresponding to different noises and environments. A convolutional recurrent network (CRN) is trained to estimate the real and imaginary spectrograms of the canceling signal from the reference signal so that the corresponding anti-noise can eliminate or attenuate the primary noise in the ANC system. Large-scale multi-condition training is employed to achieve good generalization and robustness against a variety of noises. The deep ANC method can be trained to achieve active noise cancellation no matter whether the reference signal is noise or noisy speech. In addition, a delay-compensated strategy is introduced to solve the potential latency problem of ANC systems. Experimental results show that deep ANC is effective for wideband noise reduction and generalizes well to untrained noises. Moreover, the proposed method can achieve ANC within a quiet zone and is robust against variations in reference signals.

Nonlinear Active Noise Control via Model-Based Approaches

Active noise control has attracted much research attention due to its several advantages over passive noise control. This paper introduces two model-based noise canceling techniques, that is, using the Moving Average (MA) model and a feedforward Neural Network (NN) to estimate the signal. The Least Mean Square (LMS) algorithm is used to minimize the error in the MA model while a backpropagation algorithm is employed to optimize the NN. Due to its advantages of good robustness and nonlinear processing, the NN is considered to be suitable for nonlinear signals. In order to reduce computational cost, the backpropagation algorithm in the NN is applied once at each time step with only one iteration. To examine the methods, two real-world problems are considered, one being engine noise and the other road traffic noise. A comparison between the two methods is carried out. Results indicate that both the MA and NN processors are effective in reducing the noises and that the NN based approach is superior over the MA model, especially for low frequency band.