FxLMS(Filtered-X Least Mean Square) 演算法是最經典常用的方法於主動噪音控制(Active Noise Control,ANC) 中，它運用在高斯環境的情況下通常會有良好的收斂成效，並且已經有大量的學者在進行研究並探討如何進一步改善它的效能，但是應用在現實世界中，ANC 系統可能會接收到來自環境中突然生成的詭異訊號，我們稱作脈衝噪音(Impulse Noise,IN)，這可能導致廣為人知的FxLMS 演算法效能出現劇烈下降或收斂失敗，翻閱過往一些文獻已經有在探討這方面的問題，但是許多提出的解決方法是取決於預先確定的參數，在本文中，對於脈衝噪音的問題是採用穩健函數來抑制，並且新提出的演算法與過去文獻所提的不同，不需要透過事前條件來定義參數，並結合嵌入式架構，達成播放音訊時，同時可以消除噪音，經由模擬所得的結果，我們將與過往的方法做比較與討論，而提出的演算法會擁有更良好的(Noise Reduction Ratio,NRR) 效能。;The Filtered-X Least Mean Square(FxLMS) is the most classic and commonly used method for Active Noise Control(ANC), which usually has a good convergence effect in a Gaussian environment.Therefore, there have been already a lot of scholars conducting research and discussing how to improve its performance. However,in real world the ANC system may receive a special signal suddenly generated from the environment and we called the impulsive noise(IN),which can result in degraded of performance or convergence failure to the well-known FxLMS algorithm. Some previous literatures have been taking into account this problem, but many of the proposed solutions depend on the setup of pre-determined parameters. In this article,the IN problem is dealt with robust function and the newly proposed algorithm is different from those proposed in the previous literature. It does not need to define parameters through prior conditions, and it can eliminate noise while playing audio.The results obtained through simulations are compared and discussed with the previous methods and show that the proposed has better noise reduction ratio(NRR) performance.
