二维DOA估计中麦克风阵列优化设计

在DOA估计中，往往事先假设麦克风阵列的结构，然后通过改进DOA估计方法来提高定位精度，忽视了麦克风的摆放位置对 DOA 估计性能的影响。基于此，针对二维 DOA 估计，提出改进遗传优化算法，将空时滤波器系数和麦克风阵列结构分开，构造由二维 MUSIC空间谱函数欧式距离和优化后阵元个数共为变量的适应度函数，以 DOA 估计精度为停止条件，对均匀矩形阵、均匀圆形阵和均匀同心圆阵开展优化设计。仿真结果表明，采用所提方法优化后的阵列取得了较好的DOA估计性能。

Microphone array optimization design for two-dimensional DOA estimation

The studies usually focus on how to give an effective method to improve position accuracy with the known-microphone array configurations for the DOA estimation. However, the position of the microphone influences the overall performance of DOA estimation. An improved genetic optimization algorithm is thus proposed for two-dimensional DOA estimation. Its fitness function consists of the Euclidean distance of the 2-D MUSIC spatial spectrum function and microphone numbers after optimization, which separates the space-time filter coefficients from microphone array configuration. And DOA estimation precision is adopted as stop condition of genetic algorithm. Uniform rectangular array, uniform circular array and uniform concentric circular array are used to carry out optimization design. The simulation results show that effective DOA estimation performances are obtained after optimization with the proposed method.