Layout Optimization of Cooperative Distributed Microphone Arrays Based on Estimation of Source Separation Performance

Abstract

[abstFig src='/00290001/08.jpg' width='300' text='Optimizing robot positions for source separation' ] The active audition method presented here improves source separation performance by moving multiple mobile robots to optimal positions. One advantage of using multiple mobile robots that each has a microphone array is that each robot can work independently or as part of a big reconfigurable array. To determine optimal layout of the robots, we must be able to predict source separation performance from source position information because actual source signals are unknown and actual separation performance cannot be calculated. Our method thus simulates delay-and-sum beamforming from a possible layout to calculate gain theoretically, i.e., the expected ratio of a target sound source to other sound sources in the corresponding separated signal. Robots are moved into the layout with the highest average gain over target sources. Experimental results showed that our method improved the harmonic mean of signal-to-distortion ratios (SDRs) by 5.5 dB in simulation and by 3.5 dB in a real environment.

Placement Planning for Sound Source Tracking in Active Drone Audition

“…Not utilizing activity • Source localization (direction) [45][46][47] • Source localization (location) [14,15] • Source localization (direction) [5,48] • Source localization (location) [10][11][12]24] Utilizing activity • Source localization (direction) [37,38,40] • Sound source separation [49] • Source localization (location) [This paper]…”