DOI QR코드

DOI QR Code

Nonlinear Noise Attenuator by Adaptive Wiener Filter with Neural Network

신경망 구조의 적응 Wiener 필터를 이용한 비선형 잡음감쇠기

  • Haeng-Woo Lee (Dept. of Information Communication Engineering, Namseoul University)
  • 이행우 (남서울대학교 지능정보통신공학과)
  • Received : 2022.11.24
  • Accepted : 2023.02.17
  • Published : 2023.02.28

Abstract

This paper studied a method of attenuating nonlinear noise using a Wiener filter of a neural network structure in an acoustic noise attenuator. This system improves nonlinear noise attenuation performance with a deep learning algorithm using a neural network Wiener filter instead of using a conventional adaptive filter. A voice is estimated from a single input voice signal containing nonlinear noise using a 128-neuron, 8-neuron hidden layer and an error back propagation algorithm. In this study, a simulation program using the Keras library was written and a simulation was performed to verify the attenuation performance for nonlinear noise. As a result of the simulation, it can be seen that the noise attenuation performance of this system is significantly improved when the FNN filter is used instead of the Wiener filter even when nonlinear noise is included. This is because the complex structure of the FNN filter expresses any type of nonlinear characteristics well.

본 논문은 음향잡음감쇠기에서 신경망 구조의 Wiener 필터를 이용하여 비선형 잡음을 감쇠시키는 방법에 대하여 연구하였다. 이 시스템은 기존의 적응필터를 이용하는 대신 신경망 위너필터를 이용한 심층학습 알고리즘으로 비선형 잡음감쇠 성능을 개선한다. 128-neuron, 8-neuron 은닉층과 오차 역전파(back propagation) 알고리즘을 이용하여 비선형 잡음이 포함된 단일입력 음성신호로부터 음성을 추정한다. 본 연구에서 비선형 잡음에 대한 감쇠 성능을 검증하기 위하여 Keras 라이브러리를 사용한 시뮬레이션 프로그램을 작성하고 모의실험을 수행하였다. 모의실험 결과, 본 시스템은 비선형 잡음이 포함되어 있는 경우에도 위너필터 대신 FNN 필터를 사용하면 잡음감쇠 성능이 상당히 개선되는 것을 볼 수 있다. 이는 FNN 필터의 복잡한 구조가 어떤 형태의 비선형 특성도 잘 표현하기 때문이다.

Keywords

Acknowledgement

이 논문은 2022년도 남서울대학교 학술연구비 지원에 의해 연구되었음.

References

  1. M. R. Sambur, "Adaptive noise canceling for speech signals," IEEE Trans. Acoust., Speech, Signal Processing, vol. ASSP-26, Oct. 1978, pp. 419-423.  https://doi.org/10.1109/TASSP.1978.1163137
  2. K. Lashkari, "A novel Volterra-Wiener model for equalization of loudspeaker distortions," In 2006 IEEE international conference on acoustics speech and signal processing proceedings, Toulouse, France, 2006, pp. V. 
  3. X. Guo, Y. Li, J. Jiang, C. Dong, S. Du, and L. Tan, "Sparse modeling of nonlinear secondary path for nonlinear active noise control," IEEE Transactions on Instrumentation and Measurement, vol. 67, no. 3, 2018, pp. 482-496.  https://doi.org/10.1109/tim.2017.2781992
  4. L. Tan and J. Jiang, "Adaptive Volterra filters for active control of nonlinear noise processes," IEEE Trans. Signal Processing, vol. 49, no. 8, Aug. 2001, pp. 1667-1676.  https://doi.org/10.1109/78.934136
  5. R. Napoli and L. Piroddi, "Nonlinear active noise control with NARX models," IEEE Transactions on Audio, Speech and Language Processing, vol. 18, no. 2, Feb. 2010, pp. 286-295.  https://doi.org/10.1109/TASL.2009.2025798
  6. S. Ghasemi, R. Kamil, and M. H. Marhaban, "Nonlinear THF-FXLMS algorithm for active noise control with loudspeaker nonlinearity," Asian Journal of Control, vol. 18, no. 3, 2016, pp. 502-513.  https://doi.org/10.1002/asjc.1140
  7. S. M. Kuo and H.-T. Wu, "Nonlinear adaptive bilinear filters for active noise control systems," IEEE Transactions on Circuits and Systems, vol. 52, no. 3, 2005, pp. 617-624.  https://doi.org/10.1109/TCSI.2004.842429
  8. D. P. Das and G. Panda, "Active mitigation of nonlinear noise processes using a novel filtered-s LMS algorithm," IEEE Transactions on Speech and Audio Processing, vol. 12, 2004, pp. 313-322.  https://doi.org/10.1109/TSA.2003.822741
  9. O. J. Tobias and R. Seara, "Leaky-FXLMS algorithm: stochastic analysis for Gaussian data and secondary path modeling error," IEEE Transactions on Speech and Audio Processing, vol. 13, no. 6, 2005, pp. 1217-1230.  https://doi.org/10.1109/TSA.2005.852018
  10. N. V. George and G. Panda, "Advances in active noise control: A survey, with emphasis on recent nonlinear techniques," Signal Processing, vol. 93, 2013, pp. 363-377.  https://doi.org/10.1016/j.sigpro.2012.08.013
  11. S. D. Snyder and N. Tanaka, "Active control of vibration using a neural network," IEEE Transactions on Neural Networks, vol. 6, no. 4, 1995, pp. 819-828.  https://doi.org/10.1109/72.392246
  12. M. Bouchard, B. Paillard, and C. T. Le Dinh, "Improved training of neural networks for the nonlinear active control of sound and vibration," IEEE Transactions on Neural Networks, vol. 10, no. 2, Mar. 1999, pp. 391-401.  https://doi.org/10.1109/72.750568
  13. C. Y. Chang and F. B. Luoh, "Enhancement of active noise control using neural-based filtered-x algorithm," Journal of Sound and Vibration, vol. 305, no. 1-2, 2007, pp. 348-356.  https://doi.org/10.1016/j.jsv.2007.04.007
  14. T. Krukowicz, "Active noise control algorithm based on a neural network and nonlinear input-output system identification model," Archives of Acoustics, vol. 35, no. 2, 2010, pp. 191-202.  https://doi.org/10.2478/v10168-010-0018-0
  15. G. Panda and D. P. Das, "Functional link artificial neural network for active control of nonlinear noise processes," In 2003 international workshop on acoustic echo and noise control, Kyoto, Japan, 2003, pp. 163-166. 
  16. J. Schmidhuber, "Deep learning in neural networks: An overview," Neural Networks, vol. 61, 2015, pp. 85-117.  https://doi.org/10.1016/j.neunet.2014.09.003
  17. J. S. Choi, "Speech and Noise Recognition System by Neural Network," Journal of the Korea Institute of Electronic Communication Sciences, vol. 5, Aug. 2010, pp. 357-362. 
  18. H. W. Lee, "Optimization of the kernel size in CNN noise attenuator," Journal of the Korea Institute of Electronic Communication Sciences, vol. 15, Dec. 2020, pp. 987-994.  https://doi.org/10.13067/JKIECS.2020.15.6.987
  19. H. W. Lee, "Optimization of the number of filter in CNN noise attenuator," Journal of the Korea Institute of Electronic Communication Sciences, vol. 16, Aug. 2021, pp. 625-632. https://doi.org/10.13067/JKIECS.2021.16.4.625