Title:Pouring by Feel: An Analysis of Tactile and Proprioceptive Sensing for Accurate Pouring

Abstract:As service robots begin to be deployed to assist humans, it is important for them to be able to perform a skill as ubiquitous as pouring. Specifically, we focus on the task of pouring an exact amount of water without any environmental instrumentation, that is, using only the robot's own sensors to perform this task in a general way robustly. In our approach we use a simple PID controller which uses the measured change in weight of the held container to supervise the pour. Unlike previous methods which use specialized force-torque sensors at the robot wrist, we use our robot joint torque sensors and investigate the added benefit of tactile sensors at the fingertips. We train three estimators from data which regress the poured weight out of the source container and show that we can accurately pour within 10 ml of the target on average while being robust enough to pour at novel locations and with different grasps on the source container.