A Robotic Skin-Based Approach to Soft Robotics
Date of Award
Doctor of Philosophy (PhD)
Mechanical Engineering & Materials Science (ENAS)
Typical robots are designed to achieve a single function in a controlled environment and lack the ability to generalize to new tasks. In their quest to build more capable robots, engineers have explored many avenues, including artificial intelligence, reconfigurable robots, and leveraging deformable materials that naturally absorb impacts and conform to objects. In this dissertation, I build upon these latter two bodies of work and introduce reconfigurable robotic skins, working toward the long-term vision of general-purpose robots. By applying the skins to different objects, new functionalities can be obtained, and this layer-based approach to robot design allows engineers to endow unused surfaces with desired actuation, sensing, and switchable structural properties. My focus on robotic skins has yielded several technical results, including skins that can be applied to the surface of different deformable bodies to create robots, morphing robots that change shape to overcome obstacles or change gaits to operate in new environments, and jamming skins that support objects from their surface. Combining these foundational ideas, I then demonstrate variable-stiffness robotic skins that can move on their own, in addition to locking into a desired shape. Finally, I have used the robotic skin concept to devise a new way to measure the shape of soft structures from their surface, using sensorized stretchable circuits. The reconfigurable robotic skins introduced herein could find application in fields such as healthcare and space exploration, serving as continuum manipulators, supportive exosuits, and smart textiles. Collectively, my work points toward new types of robots, where wide-ranging functionality is embedded in reconfigurable sheets.
Shah, Dylan, "A Robotic Skin-Based Approach to Soft Robotics" (2022). Yale Graduate School of Arts and Sciences Dissertations. 520.