WISARD: Weight Informing Soft Artificial Robotic Dermis

As robots are rapidly approaching the status of partners rather than tools, nonverbal communication becomes increasingly important to ensure smooth operation. In this paper, we focus on a scenario where a person receives an object of unknown weight from a collaborative robot. As humans, we are used to reading the body language of people around us. When a human lifts an object and hands it to us, we therefore expect to see a suitable reaction in the person's effort and strain. This nonverbal signal relays information to us about the approximate weight of the object, which we use for applying an appropriate force to smoothly receive the object. Inspired by this nonverbal collaborative task coordination in humans, we propose a concept for a soft robotic addition to a collaborative robot arm. WISARD (Weight Informing Soft Artificial Robotic Dermis) emulates tensed muscles to visually inform a human receiver of the approximate payload weight. Using built-in sensors of the robot arm, inflation of the pouches on the WISARD is automatically correlated with the weight of the lifted object. WISARD was tested in a robot-human handover experiment to determine if it leads to an increase in physical performance. Results indicate that subjects can use the WISARD to predict the weight of objects lifted by the robot arm and react appropriately when receiving objects of unknown weight. The most noticeable improvement was found for the reception of the heavy object. In an online video survey, we also found WISARD to increase the perceived warmth of the robot arm. The study proposes a novel use of pneumatically actuated soft robotics technology as a means to augment traditional robots with nonverbal signalling capabilities to improve human-robot collaboration.