聲音來源判別是機器人一個重要的功能如果能夠準確的判別聲音來源就可以增加機器人所能夠提供的功能。以災難反應機器人或室內服務機器人為例，透過聲音來源方向的判別，可以讓機器人轉向音源，通常是需要救援或是服務的人，並朝向音源方向移動。本論文使用麥克風陣列以判別聲音來源的方向與角度，並用到達時間差法(Time Difference Of Arrival, TDOA)與波束成形法(Beamforming)以計算聲音來源角度，並比較兩種方法的結果的誤差與使用的時間。比較結果顯示波束成型法所需計算時間遠多於到達時間差法。目前使用三個麥克風為主要的陣列，實驗在160 cm x 160 cm以及 240 cm x 240 cm 的範圍內，以手機播放聲音來當作聲源加以測試。使用TDOA法，前者之誤差角度在0.2%~8.2%區間，後者之誤差角度在0.8%~4.75%區間。使用波束成型法所計算之誤差角度在0~10%之間與使用TDOA法比較起來沒有太大差別。

Identifying the direction of some sound source is one of the important issues in developing robots. Taking a disaster response robot or a home-service robot as an example, as the direction of the sound source, which would be a person, is determined, it could turn its direction to face and move toward the person.

This research used the Time Difference of Arrival (TDOA) method and the Beamforming method, separately and compared the calculation results of the sound source angle and the computation time. We found that TDOA saves more computation time than the latter. The receiver of the sound source is a microphone array. Three microphones are used as the main array. The experiments were conducted in the range of 160cm x 160 cm and 240cm x 240 cm. A mobile phone kept broadcasting music served as the sound source. The error angle of the former is between 0.2% and 8.2%, and the error angle of the latter is between 0.8% and 4.75%.The error angle calculated by the Beamforming method is between 0 and 10%, which is not much different from the TDOA method.

[1] C. Knapp, G. Carter, "The Generalized Correlation Method For Estimation of
Time Delay", IEEE Transactions on Acoustics, 1976, Vol. 24, no. 4,
pp. 1320 - 1323
[2] Daniele Salvati, Carlo Drioli, Gian Luca Foresti, "Sound Source and
Microphone Localization from Acoustic Impulse Responses", IEEE Signal Processing Letters, 2016, Vol. 23, no. 10, pp. 1459 – 1463
[3] Dongsuk Yook, Taewoo Lee, Youngkyu Cho ,"Fast Sound Source
Localization Using Two-Level Search Space Clustering", IEEE Transactions on Cybernetics, 2015, Vol. 46, no. 1, pp. 20 – 26
[4] Hung-Yan Gu, Shan-Siang Yang, "A Sound-Source Localization System
Using Three-Microphone Array and Cross-power Spectrum Phase", Proc. of International Conference on Machine Learning, 2012, pp. 15-17, Xian, China
[5] Jiri Tuma, Patrik Janecka, Milan Vala and Lukas Richter, "Sound
Source Localization", Proc. of International Carpathian Control Conference, 2012, pp. 28-31, High Tatras, Slovakia
[6] Jacob Benesty, Jingdong Chen, Yiteng Huang, “Microphone Array Signal
Processing”, Springer Verlag, 2008, Chapter 03, Berlin, Germany
[7] Jacob Benesty, Jingdong Chen, Yiteng Huang, “Microphone Array Signal
Processing”, Springer Verlag, 2008, Chapter09, Berlin, Germany
[8] Yanwen Li, Huawei Chen, "Reverberation Robust Feature Extraction for
Sound Source Localization Using a Small-Sized Microphone Array",
IEEE Sensors Journal, 2017, Vol. 17, no. 19, pp.6331 – 6339
[9] Nitesh Kumar Chaudhary, Subir Verma, Anshu Aditya, "Sound Source
Localization using GCC-PHAT with TDOA Estimation", Journal of Basic and Applied Engineering Research, 2014, Vol. 1, no. 11, pp. 54-58