具有快速收歛之主動噪音控制系統

Fast Convergence Active Noise Control Systems

徐榮昌(ANTONIUS SISWANTO)

指導教授 ： 張政元

中原大學/電機資訊學院/電機工程學系/博士(2022年)

https://doi.org/10.6840/cycu202201397

摘要

主動噪音控制 (active noise control, ANC)系統使用FXLMS (filtered-x least mean square)算法更新自適應濾波器來減少不需要的寬頻和窄頻噪音，以應對不斷變化的環境噪音，降噪性能會受到FXLMS算法收斂速度的影響，以使自適應濾波器收斂到其最佳狀態，本論文的重點為快速收斂ANC系統的分析與開發。
傳統的 ANC 系統在降低具有較大頻譜動態範圍的寬頻噪音方面收斂緩慢，本論文開發了一種並行形式的寬頻ANC系統，該系統將自適應濾波器分離為一組具有並行結構的單一權重自適應濾波器以提高收斂速度。此外本論文還分析了窄頻噪音的頻率對於窄頻ANC系統中使用FXLMS算法收斂速度的影響，提出了基於頻率依賴性的快速收斂濾波器長度，並表明了它是實現快速收斂的關鍵因素。
ANC系統可以集成到現有的音頻系統中，以降低整體成本，然而傳統的音頻集成窄頻反饋 ANC (audio-integrated narrowband feedback active noise control, ANFANC)系統以單一取樣率運行，這導致 ANC系統在音頻系統所需的高取樣率的運行中需要極大的運算量，本論文還提出了多速率ANFANC系統，該系統在保持快速收斂的同時僅需更低的運算量，呈現了電腦模擬和即時實驗以支持所提出系統的有效性。

關鍵字

主動噪音控制

並列摘要

Active noise control (ANC) systems reduce unwanted broadband and narrowband noises by using adaptive filters updated by the filtered-x least mean square (FXLMS) algorithm to cope with the changing noises and environments. The noise reduction performance is influenced by the convergence rate of the FXLMS algorithm to converge the adaptive filters into their optimal state. This dissertation focuses on the analysis and development of fast convergence ANC systems.
The conventional ANC systems are subjected to slow convergence in reducing broadband noises with large spectral dynamic range. This dissertation develops the parallel-form broadband ANC system that separates the adaptive filter into a bank of single weight adaptive filters in parallel structure that is capable to enhance the convergence rate. Besides, this dissertation also analyzes the effect of narrowband noises’ frequencies on the convergence rate of FXLMS algorithm used in the narrowband ANC systems. The fast convergence filter length based on the frequency dependence is proposed, and shown to be a crucial factor in achieving fast convergence.
The ANC system can be integrated into the existing audio systems to reduce the overall cost. However, the conventional audio-integrated narrowband feedback ANC (ANFANC) systems operate on the single sampling rate that causes the ANC system requires huge computational load in high sampling rate operation required by the audio system. This dissertation also proposes the multi-rate ANFANC system that requires less computational load while maintaining fast convergence. The computer simulations and real-time experiments are presented to support the effectiveness of the proposed systems.

並列關鍵字

active noise control

參考文獻

[1] World Health Organization, “Burden of Disease from Environmental Noise-Quantification of Healthy Life Years Lost in Europe,” 2011. Accessed: May 20, 2022 [Online]. Available: https://www.euro.who.int/__data/assets/pdf_file/0008/136466/

e94888.pdf

[2] F. Khajenasiri, A. Zamanian, Z. Zamanian, “The Effect of Exposure to High Noise Levels on the Performance and Rate of Error in Manual Activities,” Electron. Physician, vol.8, no. 3, pp. 2088-2093, Mar. 2016.

[3] Department for Environment, Food, & Rural Affairs of United Kingdom. “Noise Pollution-Economic Analysis.” https://www.gov.uk/guidance/noise-pollution-economic-analysis (accessed May 20, 2022)

[4] T. K. Swinburn, M. S. Hammer, and R. L. Neitzel, “Valuing Quiet: An Economic Assessment of U.S. Environmental Noise as a Cardiovascular Health Hazard,” Am. J. Prev. Med., vol. 49, no. 3, pp. 345-353, Sept. 2015.