Listen in Domestic Environments with Robotic Acoustic Localization Method

This paper describes the work by using audio as a monitoring means to understand domestic environments as well as improving interaction capabilities between human and robot for monitoring task. The primary goal of this work is to provide an auditory awareness of the monitored site to a remote user; specifically the emphasis is put on sound localization which has been widely studied with the aim of understanding the mechanism underlying the biological systems. Nevertheless, in robotics the use of auditory cues has been rather limited. Thus we present a statistical approach with interaural cues which is able to estimate sound source directions reliably. It also has been shown that this method work even well when implementing interpolation scheme into it with an enhanced accuracy. Experimental results show that accurate acoustic localization can be achieved using our method