Design Principles for Improving Precision and Dexterity of Soft Robotic Hands
Teeple, Clark Benjamin
MetadataShow full item record
CitationTeeple, Clark Benjamin. 2021. Design Principles for Improving Precision and Dexterity of Soft Robotic Hands. Doctoral dissertation, Harvard University Graduate School of Arts and Sciences.
AbstractAs robots move more and more into the real world, they need to be equipped to interact with a wide array of objects and environment features while maintaining a gentle touch. Soft robotic end effectors enable passive adaptability, but give up strength and precision in return. From simple grasping to in-hand manipulation, this thesis explores how the design of soft robotic hands influences a robot's overall manipulation capabilities. We begin with a foray into finger design, developing pneumatically-actuated soft fingers capable of robust precision grasping and power grasping via two independently-actuated bending segments. This new finger design forms the basis of a further investigation into the role of gripper compliance in grasping and manipulation of thin, flexible materials, where we find that vertical, lateral, and rotation compliance all play a role in minimizing damage. With a similar goal of gentle interaction, we then shift focus to dexterous in-hand manipulation, exploring how distributing controlled degrees of freedom into various parts of the hand contributes to overall dexterity. We first develop a dexterous soft finger design and soft hand platform capable of moving objects within the hand. Using this new hand platform, we show that active control of the arrangement of digits affects the categories of objects that can be successfully manipulated. We also show that controlling the interaction between objects and the palm (via the palm's frictional properties and location relative to fingers) enables greater grasp stability and expanded access to different motion primitives. Finally, we collect these results into a set of application-specific design principles which can be used to inform the design of soft hands with dexterity tuned for the particular application.
Citable link to this pagehttps://nrs.harvard.edu/URN-3:HUL.INSTREPOS:37371111
- FAS Theses and Dissertations