Toward Learning to Press Doorbell Buttons

To function in human-inhabited environments a robot must be able to press buttons. There are literally thousands of different buttons, which produce various types of feedback when pressed. This work focuses on doorbell buttons, which provide auditory feedback. Our robot learned to predict if a specific pushing movement would press a doorbell button and produce a sound. The robot explored different buttons with random pushing behaviors and perceived the proprioceptive, tactile, and acoustic outcomes of these behaviors. Previous related work in robotics has focused mainly on visual feedback. Thomaz (2006) proposed using social guidance to teach a robot which button it has to press. Miura et al.’s (2005) robot searched for an image template of a button in the region pointed to by a human and pressed it. Nguyen et al. (2009) used PPS-Tags to make switches more visible so that a robot can press them more easily. In contrast, the robot in our study relied exclusively on multi-modal feedback, i.e., auditory, proprioceptive and tactile feedback. In Psychology, Hauf and Ascherleben (2008) reported that 9 months old human infants already anticipate that pressing a button would produce interesting sounds or lights. E.J. Gibson (1988) suggested that humans actively apply exploratory behaviors on objects to learn their affordances.