Improvement of robot audition by interfacing sound source separation and automatic speech recognition with Missing Feature Theory | IEEE Conference Publication | IEEE Xplore