微機械陣列麥克風之 System-on-chip 設計

Abstract

此論文主要在發展一結合微機電技術和陣列訊號處裡之微小的陣列麥克風。 將微機械電容式麥克風和電子元件以及陣列訊號處理之系統整合在單一晶片。 在微機械電容式麥克風裡，探討微機械電容式麥克風之靜態和動態的行為分析。 靜態行為分析主要在探討麥克風振膜之非線性行為和求得麥克風最佳之性能，在此利用有限差分方法來模擬。 而動態行為分析主要再探討麥克風的靈敏度，在此利用機電聲類比將麥克風之結構等效成電子元件來做模擬。 在陣列訊號處理部分，將探討陣列訊號處理的基本理論以及方向估測和波束形成之設計和模擬。 另外，針對微機械陣列麥克風之結構，提出一改良之方法，用以增進麥克風之指向性。 此方法是利用最小平方法去設計每個麥克風所需經過的濾波器，使得陣列麥克風之空間波束在每個頻率都能近似於一理想的窄波束。 此波束樣板是由MUSIC方法來求得，因為MUSIC演算法擁有很窄的波束去判斷聲源方向。 此濾波器可以利用Biquad電路來加以實現。 則此一類比電路能和微機械陣列麥克風以及阻抗匹配電路，訊號放大電路能夠整合於單一晶片上。

A miniature microphone array is developed which combines the array signal processing and MEMS technology. In the array signal processing, the fundamental array theories, which include the direction of arrival (DOA) estimation and beamforming design, will be introduced. Delay-sum is the most popular algorithm to estimate DOA. However, the delay-sum incurs a problem of fractional delay. The Lagrange interpolation is used to overcome this problem. In addition to delay-sum, some high-resolution algorithms are discussed and simulated. These algorithms contain the MVDR, AR and MUSIC DOA estimation. The beamformer is utilized for improving the directionality of the array. Besides, the beamformer can enhance the signals of the array. Here we have to design the filters to materialize the beamformer. A least square beam design by MUSIC template is proposed to design the filters. Using the high resolution characteristic of the MUSIC has a great improvement of the directionality. In order to design a MEMS microphone array, firstly the MEMS microphone has to be analyzed. The MEMS microphone is investigated into two parts: the quasi-static properties and dynamic behaviors. The quasi-static is used to analyze the nonlinear deflection of the diaphragm of the microphone. Here the finite difference method (FDM) is exploited to analyze the quasi-static behavior. The dynamic behaviors are analyzed for small variations of the diaphragm of the microphone. These variations induce to alternating current signals. A common way to analyze the dynamic behaviors is to use the analogy. The analogy will be show to simulate the frequency response of the microphone. In the following, aforementioned beamforming filters will be achieved by analog circuits. Finally, the system-on-chip design of a MEMS microphone array will be realized by integrating the MEMS microphone and beamforming filters in the same chip.