University of Cambridge, UK
Lab webpage: https://birlab.org/
The granular jamming-based universal gripper is able to perform stiffness transition for grasping arbitrarily shaped objects. The conventional universal gripper perform the grasping with a force feedback-free approach, thus making it easy to crush the object (especially the brittle and fragile one). In this paper we propose a sensorized universal gripper with force sensing and contact localization ability when it is pressed against an object, allowing it to achieve adaptive grasping and safe interaction with humans. Traditional soft sensors with thermoplastic elastomer for sensory strain information purposes are susceptible to break from large mechanical deformation when sensing unstructured objects. We address the challenge by using a self-healing 4×4 sensor array fabricated from conductive hydrogels. The sensor is able to heal from microscopic damages without external stimuli under room temperature, making the gripper promising for long-life use. A machine learning framework has been proposed to address the innate hysteresis and drift of the hydrogel soft sensors. The physical experimentation has demonstrated that the sensorized UG realizes force feedback and contact localization during grasp task with <3% and <5% error rates, respectively. The soft sensor retains sensory effectiveness and accuracy after self-healing from fatal damage.
Huijiang Wang is a PhD student at the University of Cambridge, UK as a Marie Curie Fellow under the EU Horizon 2020 SMART Innovative Training Network. His research interests include soft robotics, tactile sensing, machine learning and bio-inspired algorithms.