自适应多通道双L型麦克风阵列声源定位算法

Adaptive Multichannel Double L-Shaped Microphone Array Sound Source Localization Algorithm

针对室内环境中麦克风阵列声源定位算法抗噪声和混响能力不强、定位估计精度低等问题,提出了一种基于自适应多通道双L型麦克风阵列的室内声源定位算法。在单源多元混响模型下,采用归一化多通道频域最小均方算法,估计了不同阵元的冲激响应并计算了时延;由惩罚函数对频谱能量进行了修正,提高了抗噪声能力、避免了信道估计恶化;采用滤波器长度限制法,提高了抗混响能力;利用双L型麦克风阵列结构进行室内声源定位,提高了定位精度。实验表明,在室内环境下,文中算法具有较强的声源定位性能,且能满足实际要求。

Aiming at the lack of ability to resist noise and reverberation and the location estimation low accuracy for the sound source localization method based on microphone array in the indoor sound environments, a sound source localization algorithm based on adaptive multichannel double L-shaped microphone array was proposed. In this proposed algorithm, normalized multichannel frequency-domain least mean square algorithm was used to estimate the impulse responses of different array elements and acquire the time delay under condition of the single source and multiple reverberation model, the spectrum energy was modified by penalty function to improve the ability to suppress noise and avoid channel estimation deterioration, and the method on limited length of the filter was employed for improving the ability to suppress reverberation, as well as double L-shaped microphone array structure was used to improve the localization accuracy in the indoor sound environments. The simulation results show that the proposed algorithm has higher performance of sound source localization in indoor environment and meets the actual demand.