適用於高噪音環境之主動式噪音消除方法

Abstract

本計畫著眼於開發可商業化的主動噪音消除(Active noise cancellation)相關訊號處理技術。主動式噪音 消除的技術，主要是偵測背景噪音的大小以及相位，然後應用適應性訊號處理(Adaptive filter)以及訊 號預測(Signal prediction)的方式，製造一個相對於背景噪音的反相波給使用者。如此一來，背景雜訊 就會被消除，可以讓使用者有更舒適安靜的環境。目前，已經有主動式噪音消除技術的主要是利用 特殊應用數位電路(Application specific integrated circuit, ASIC)來實現及時(real-time)的訊號處理。在 此計畫中，希望可以開發出應用在高速數位訊號處理器為主體來運算主動噪音消除的所需要的數位 訊號處理技術。並且能夠以低運算複雜度以及高噪音消除能力的演算法為原則，來進行技術開發。 但是，此技術也必須要能夠容忍不同麥克風之間，因生產偏差問題造成的頻率響應不符合等的差異， 才能夠達到商業化量產的目標。

This project aims to develop a commercializable signal processing technique for the active noise cancellation (ANC). This technique mainly estimates the amplitude and phase of the background noise and utilizes the knowledge of the adaptive filtering and signal prediction to generate a signal wave opposing to the background noise. As a result, the background noise can be cancelled and allow the user to have a more comfortable and quieter environment. Nowadays, the ANC technique has been implemented in the ASIC in order to achieve the real-time signal processing. However, the goal in this project is to develop a signal processing technique based on a high speed digital signal processor so as to provide the computational power for ANC in the principle of low computational complexity and high noise cancellation capability. Furthermore, in order to achieve the goal of being mass-producible, this technique also needs to handle the discrepancy of frequency responses between different microphones due to the mass-production process.