64ch Design and Implementation of Omni-Directional Ball Microphone Array,, Simon THOMPSON,, Yoko SASAKI, Mitsutaka KABASAWA, Simon THOMPSON, Satoshi KAGAMI, Kyoichi ORO National Institute of Advanced Industrial Science and Technology Kansai Electric Power Co., Inc. y-sasaki@aist.go.jp Abstract This paper presents a microphone array design and the evaluation result of the developed microphone array. We propose an evaluation index of directional characteristic of Delay and Sum BeamForming to optimize microphone array design. Using beamforming simulation, we obtain a microphone arrangement which minimizes sidelobes, and improves the basic performance of beamforming. It has 64 microphones in a 35mm diameter ball designed to mount on a mobile robot and omni-directional directivity in azimuth and elevation. The performance of the proposed microphone array is verified in different real environments. Experimental results of sound localization show the effectiveness of the array in some challenging environment and its robustness for different pressure sound sources to cover larger areas. 1 [1, 3, 4, 6] (Delay and Sum BeamFoming, DSBF) DSBF 12ch [7, 9]DSBF DSBF [5] Griffith-Jim SN MUltiple SIgnal Classification (MUSIC) DSBF 2 DSBF 3
2.1 C (l, θ, ϕ) C i (i = 1, 2,, M) L i i V s τ i (l, θ, ϕ) = l L i(l, θ, ϕ) V s (1) t i x i (t) C DSBF (2) s c (t) = 1 M M x i (t + τ i ) (2) i=1 ( ) 2.2 DSBF 2 IMs : ISmax : -12dB Figure 1 db ω ω L, ω H N ω L ω H α a, b a IMs b ISmax 3 3.1 r [2] DSBF ρ ρ 1/ cos(r) (4) 4 ρ(r) =.328.117.496.117 cos ( r + )4 cos ( r )3 cos ( r + )2 cos ( r R R R R ).122 (4) (4) 1. 2. 3. 2 ( /) (4) (3) 3.2 Figure 1: Directivity pattern of DSBF ω H ω H Eva = α 1 1 IMs ω 1 1 ISmax ω (3) a N b N ω=ω L ω=ω L DSBF SN 6 8 (4) 1 4
:.2,.3,.4,.6,.8, 1. [m] : 3, 5, 7, 9, 11, 13, 15, 17 3842 2 (3) Eva Figure 2 a) b) c) a) b) SN Figure 2 M = 64 36 mm 4.1 (IMs, ISmax) = (15%, -14dB) 3.3 M = 64 36 mm 5 3Hz 1Hz Figure 3 7 Figure 3 α = 1.5, 1.7 1 35mm 51 35mm 5 C6 8 2 6 5 14 ( 48 65 76 84 9 95 1 13 [deg] ) Figure 3: Evaluation Index Distribution (No Constraint) Figure 4 3 35mm 5 15mm 12 2 (IMs, ISmax) = (9.1%, -15.1dB) 4.1 Figure 4: Evaluation Index Distribution (Fixed outside) Figure 5 (θ, ϕ) = (18, ) 4 5
a) Mainlobe size b) Sidelobe gain c) Performance Boundary Figure 2: Consideration of Array Size and Number of Microphones ϕ = ϕ = 9 a) 1 Hz a) Microphones Drawing b) Fixture Drawing b) 2 Hz Figure 5: Directivity Pattern of Proposed Microphone Array 4.1 64 Figure 6 CAD a) b) 3 2 mm -2 9 () c), d) Segway RMP2 ATV 2 θ = θ = 9 c) Developed Microphone Array d) Mobile Robot Figure 6: Developed Microphone Array Table 1 16kHz Table 1: Spec. of The Microphone Array Board Microphone Primo EM1PT Num. of Channels 64 Sampling Frequency Resolution Amplifier 8, 16, 32, 48 [khz] 16 [bit] AK4563A (Programmable Gain Amp.) Interface USB 2. Power Supply +5 [v] 6
4.2 DSBF FBS(Frequency Band Selection) [8] 124 (64msec) 4.2.1 / (YAMAHA11III) SNR 1dB 3m, 9m, 15, 3, 45deg Figure 7, 8 / / 3.6deg angle error [deg] 4 3.5 3 2.5 2 1.5 1.5 angle azimuth elevation 5 1 15 2 25 3 35 4 45 elevation angle [deg] Figure 7: Average Sound Localization Error in Static Condition (distance=3m) angle error [deg] 4 3.5 3 2.5 2 1.5 1.5 angle azimuth elevation 5 1 15 2 25 3 35 4 45 elevation angle [deg] Figure 8: Average Sound Localization Error in Static Condition (distance=9m) 4.2.2 Figure 9 2 2[m] 6 3 GPS(Garmin Geko31) [deg] NTP 1.3m/s y [m] 2 robot path 15 1 5-5 -25-2 -15-1 -5 5 1 15 x [m] Figure 9: Robot Path in the Power Substation Figure 1, 11 Figure 1 Figure 9 posa θ = 18 1m 3x6 θ = 45 18m elevation [deg] azimuth [deg] 9 6 3 315 27 225 18 135 9 45 2 4 6 8 1 12 14 time [s] Figure 1: Sound Localization Result in the Power Substation: static condition at posa Figure 11 7
3 (θ = 27 ) elevation [deg] azimuth [deg] 3 315 27 225 18 135 9 45 5 1 15 time [s] 2 25 3 Figure 11: Sound Localization Result in the Power Substation: from moving robot 5 DSBF DSBF [1] Hideki Asoh, Isao Hara, and Futoshi Asano. Tracking human speech events using a particle filter. In Proceedings of IEEE International Conference on Acoustics, Speech, Signal Processing (ICASSP 25), pp. MSP P2.6, Philadelphia, USA, 25. [2] Tomoaki Fujihara, Yoko Sasaki, Satoshi Kagami, and Hiroshi Mizoguchi. Arrangement optimization for narrow directivity and high s/n ratio beam forming microphone array. In Proceedings of the 7th Annual IEEE Conference on SENSORS (IEEE SENSORS 28), pp. 45 453, Lecce, Italy, October 28. [4] Hyun-Don Kim, Jong-Suk Choi, and Munsang Kim. Speaker localization among multi-faces in noisy environment by audio-visual integration. In Proceedings of IEEE-RAS International Conference on Robots and Automation (ICRA26), pp. 135 131, Orlando, Florida, May 26. [5] Nikolaos Mitianoudis and Mike E. Davies. Audio source separation: Solutions and problems. International Journal of Adaptive Control and Signal Processing, Vol. 18, No. 3, pp. 299 314, March 23. [6] Kazuhiro Nakadai, Hirofumi Nakajima, Masamitsu Murase, Satoshi Kaijiri, Kentaro Yamada, Yuji Hasegawa, Hiroshi G. Okuno, and Hiroshi Tsujino. Real-time tracking of multiple sound sources by integration of in-room and robot-embedded microphone arrays. In Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS-26), pp. 852 859, Beijing, China, September 26. [7] Harvey F. Silverman, William R. Patterson III, and Joshua Sachar. Factors affectiong the performance of large-aperture microphone array. The Journal of the Acoustical Society of America, Vol. 111, No. 1, pp. 2144 2157, May 22. [8] Yuki Tamai, Yoko Sasaki, Satoshi Kagami, and Hiroshi Mizoguchi. Three ring microphone array for 3d sound localization and separation for mobile robot audition. In Proceedings of 25 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS25), pp. 93 98, Edmonton, Canada, August 25. [9] E. Weinstein, K. Steele, A. Agarwal, and J. Glass. Loud: A 12-node modular microphone array and beamformer for intelligent computing spaces. Technical Report MIT-LCS-TM-642, MIT/LCS Technical Memo, April 24. [3] Carlos Toshinori Ishi, Shigeki Matsuda, Takayuki Kanda, Takatoshi Jitsuhiro, Hiroshi Ishiguro, Satoshi Nakamura, and Norihiro Hagita. Robust speech recognition system for communication robots in real environments. In Proceedings of IEEE-RAS International Conference on Humanoid Robots(HUMANOIDS26), pp. 34 345, Genova, Italy, December 26. 8