Untethered Soft Robots With Shape Memory Alloy for Dynamic Locomotion
[Thesis]
Huang, Xiaonan
Majidi, Carmel
Carnegie Mellon University
2020
153
Ph.D.
Carnegie Mellon University
2020
As an emerging class of biologically-inspired machines that are primarily composed of elastomers, fluids, and other forms of soft matter, soft robots have potential advantages over traditional piece-wise rigid machines in maneuvering through complicated environments and resisting external disturbances. Despite drastically growing interests and efforts in soft robotics, building a soft untethered fast robot remains a significant challenge. To address this, I first create a highly dynamic soft load-bearing shape memory alloy (SMA) actuator that can be actuated by miniaturized electronics. Then I build several tethered and untethered soft robots that are capable of moving at biologically relevant speed and navigating unstructured terrains with the soft SMA actuators. To further improve the speed and robustness of the robot, I integrate several soft multifunctional composite materials into the soft robots, including a thermally conductive liquid metal embedded elastomer, an autonomously electrically self-healing composite and a highly conductive silver-hydrogel composite. Next, I use several soft robotic testbeds that are capable of crawling, jumping, rolling and swimming to experimentally validate a planar discrete differential geometry based physics engine that is widely used in the computer graphics community. Finally, I adapt the physics engine to simulate multi-limbed underwater robots and use it as a computational tool to optimize the design and control parameters of an untethered dynamic frog-inspired soft robot to achieve fast and efficient locomotion.