Self-Supervised Learning of Motion-Induced Acoustic Noise Awareness in Social Robots

摘要

With the growing presence of robots in human populatedenvironments, it becomes necessary to render theirpresence natural, rather than invasive. To do that, robotsneed to make sure the acous??c noise induced bytheir mo??on does not disturb people nearby. ??n this line,this paper proposes a method that allows the robot tolearn how to control the amount of noise it produces,taking into account the environmental context and therobot’s mechanical characteris??cs. ??oncretely, the robotadapts its mo??on to a speed that allows it to produceless noise than the environment’s background noise and,hence, avoiding to disturb nearby humans. For that, beforeexecu??ng any given task in the environment, the robotlearns how much acous??c noise it produces at di??erentspeeds in that environment by gathering acous??c informa??on through a microphone. The proposed methodwas successfully validated on various environments withvarious background noises. ??n addi??on, a ?????? sensor wasinstalled on the robot in order to test the robot’s ability totrigger the noise-aware speed control procedure when aperson enters the sensor’s field of view. The use of a sucha simple sensor aims at demonstra??ng the ability of theproposed system to be deployed in minimalis??c robots,such as micro unmanned aerial vehicles.