暫態噪音聲源方位追蹤

Transient Noise Source Tracking

張哲鳴(Che-Ming Chang)

指導教授 ： 胡竹生

國立交通大學/工學院/聲音與音樂創意科技碩士學位學程/碩士(2012年)

https://doi.org/10.6842/NCTU.2012.01009

摘要

本論文提出了一套偵測暫態噪音並使用麥克風陣列追蹤暫態噪音聲源方位的方法。麥克風接收的訊號經過時域振幅刪減法處理之後，可以非常有效的消除麥克風接收到的穩態與非穩態的非暫態噪音訊號對於偵測準確度的影響，對於語音也有一定的抑制效果。使得本方法在環境不理想時也有相當可靠的辨識率，可取代關鍵字做為另一種聲音喚醒機制的選擇。本方法對低維度的矩陣做運算，只需要幾個接收訊號的音框，同時從中辨識出只存在暫態噪音的音框，並只針對這些音框進行聲源方位追蹤，這樣運算量低的特性很適合應用於即時系統上。

關鍵字

暫態噪音 ； 聲源追蹤 ； 聲源到達角度估測 ； 聲音活動偵測

並列摘要

This thesis presents a method of detecting transient noise using microphone array so the transient noise source orientation can be computed. Through the time-domain amplitude subtraction, the effect of both stationary and non-stationary noise in the microphone signal can be effectively eliminated. This includes signals such as voice. It is shown that this method is reliable when the environment is not ideal. This makes the method a better candidate to be a sound cue than key word based mechanism. This method operates on a low-dimensional matrix and needs only a few window frames of receive signals. Meanwhile the source location can also be determined from those frames. This low computational requirement makes it ideal for real-time applications.

並列關鍵字

Transient Noise ； Source Tracking ； DOA ； VAD

參考文獻

[1] S.F. Boll, “Suppression of acoustic noise in speech using spectral subtraction,” IEEE Trans. Acoust., Speech, Signal Process., vol.27, pp. 113-120, Apr. 1979.

[2] R. Talmon, I. Cohen, and S. Gannot, “Speech enhancement in transient noise environment using diffusion filtering,” Proc. 35th IEEE Internat. Conf. Acoust. Speech and Signal Process. (ICASSP-2010), Dallas, Texas, pp. 4782–4785, Mar. 2010.

[3] Wen-Jun Zeng and Xi-Lin Li, “High-Resolution Multiple Wideband and Nonstationary Source Localization With Unknown Number of Sources,” IEEE Trans. Signal Process., vol. 58, no. 6, pp. 3125–3136, 2010.

[5] J.M. Valin, F.Michaud, and J. Rouat, “Robust localization and tracking of simultaneous moving sound sources using beamforming and particle filtering.,” Robotics and Autonomous Systems Journal (Elsevier), vol. 55, no. 3, pp. 216 – 228, 2007.

[6] J.-S. Hu, M.-T. Lee, and T.-C. Wang, “Wake-Up-Word Detection for Robots Using Spatial Eigenspace Consistency and Resonant Curve Similarity,” Robotics and Automation, 2011. ICRA ’11. IEEE International Conference on, pp. 3901–3906, 2011.