近场声源方位和距离联合估计算法

Joint Estimation Algorithm for DOA and Range of Sound Source in Near-field

为提高近场麦克风阵列的声源定位精确度,对基于麦克风阵列的声源定位技术进行研究。详细分析了近场信号传播模型,并结合窄带子空间算法,对宽带ISM(Incoherent Signal-Subspace Method)算法进行改进,提出一种近场子空间声源方位和距离联合估计算法。该算法根据短时信号的频率特性,在特定频率上对信号应用子空间算法进行方向距离联合估计并对结果进行叠加,从而得到最终结果。该算法在相关信源和混响条件下均能得到较好的DOA(Direction-of-Arrival)估计效果。计算机仿真验证了算法的有效性。

The technology based on microphone array for locating sound source is studied and a near-field signal transmission model is built in detail for improving the accuracy in sound source localization of the microphone array in near-field.A joint estimation algorithm for DOA（Direction of Arrival） and range of sound source in near-field is proposed by combining this model and subspace methods and modifying the ISM（Incoherent Signal-Subspace Method） algorithm based on wideband signal.The algorithm,based on the frequency quality of the short-time signal,estimates the DOA and range of the signal in special frequencies using subspace algorithm and adds them up to obtain the final result.The algorithm shows high performance on correlated signals and reverberation noise condition.The algorithm effectiveness is verified by computer simulation.