Enhanced robot audition based on microphone array source separation with post-filter

Abstract

Abstract-We propose a system that gives a mobile robot the ability to separate simultaneous sound sources. A microphone array is used along with a real-time dedicated implementation of Geometric Source Separation and a post-filter that gives us a further reduction of interferences from other sources. We present results and comparisons for separation of multiple non-stationary speech sources combined with noise sources. The main advantage of our approach for mobile robots resides in the fact that both the frequency-domain Geometric Source Separation algorithm and the post-filter are able to adapt rapidly to new sources and non-stationarity. Separation results are presented for three simultaneous interfering speakers in the presence of noise. A reduction of log spectral distortion (LSD) and increase of signalto-noise ratio (SNR) of approximately 10 dB and 14 dB are observed.

Robotic Orientation towards Speaker for Human-Robot Interaction

“…In [20], the sources are separated from each other, in order to enhance speech recognition, and as a preamble for DOA estimation. However, it required an array of 8 microphones positioned in a cube-like manner to work, which may be impractical for service robots.…”

“…A sophisticated recording system may be able to overcome them, such as the one proposed in [17] that used a 24-microphone 1-D array for precision. However, a high amount of microphones may be impractical to carry by many of the currently-in-use service robots [20,11], such as our in-house robot, Golem, herein described.…”

Microphone array beampattern characterization for hands-free speech applications

“…Microphone array based speech acquisition is a common practice in hands-free interfaces in state-of-the-art applications such as distant speech recognition [1], speech separation [2], [3] and speaker localization [4], [5]. Multi-channel signal acquisition relies on beamforming or spatial filtering for directional discrimination, and space-time filtering of the signals in the acoustic scene [6], [7].…”

Simultaneous localization of mobile robot and multiple sound sources using microphone array

“…Under this requirement there have been several research issues currently active in the Robotics community. These issues include speaker localization [1] [5], speech separation and enhancement [2], speech recognition and natural dialog [3], and speaker identification and multi-modal interaction [4] etc. Among them, speaker localization using either biological hearing principle [5] or microphone array [1] has drawn lots of attentions for many years [6].…”