ISCA Archive Interspeech 2021
ISCA Archive Interspeech 2021

Reliable Intensity Vector Selection for Multi-Source Direction-of-Arrival Estimation Using a Single Acoustic Vector Sensor

Jianhua Geng, Sifan Wang, Juan Li, JingWei Li, Xin Lou

In the context of multi-source direction of arrival (DOA) estimation using a single acoustic vector sensor (AVS), the received signal is usually a mixture of noise, reverberation and source signals. The identification of the time-frequency (TF) bins that are dominated by the source signals can significantly improve the robustness of the DOA estimation. In this paper, a TF bin selection based DOA estimation pipeline is proposed. The proposed pipeline mainly involves three key steps: key frame identification, TF bin selection and DOA extraction. We identify the key frames by frame-wisely examining the effective rank. Subsequently, the geometric medians of the selected key frames are extracted to alleviate the impact of extreme outliers. The simulation results show that the accuracy and the robustness of the proposed pipeline outperform the state-of-the-art (SOTA) techniques.


doi: 10.21437/Interspeech.2021-375

Cite as: Geng, J., Wang, S., Li, J., Li, J., Lou, X. (2021) Reliable Intensity Vector Selection for Multi-Source Direction-of-Arrival Estimation Using a Single Acoustic Vector Sensor. Proc. Interspeech 2021, 2137-2141, doi: 10.21437/Interspeech.2021-375

@inproceedings{geng21_interspeech,
  author={Jianhua Geng and Sifan Wang and Juan Li and JingWei Li and Xin Lou},
  title={{Reliable Intensity Vector Selection for Multi-Source Direction-of-Arrival Estimation Using a Single Acoustic Vector Sensor}},
  year=2021,
  booktitle={Proc. Interspeech 2021},
  pages={2137--2141},
  doi={10.21437/Interspeech.2021-375}
}