The emerging field of soft robots has shown great potential to outperform their rigid counterparts due to the soft and safe nature and the capability of performing complex and compliant motions. Many are built, but the designs are conservative and limited to regular shapes. The widely-used fabrication method contains bulky pumps, tethered tubings, and silicone elastomer that takes hours to fully cure. Control methods for soft robots, both model-based and learning-based, are investigated, but none of them is applied to soft robots with complex shapes. In this thesis, a set of tools and methodology is developed for cables-driven soft robots, including 1) an automated design system that generates optimal cable placement designs given desired configuration, 2) a forward simulator that predicts the soft body motion under cable contractions, 3) a fabrication pipeline that is fast, low-cost, and accessible to non-experts, and 4) various algorithms to solve inverse kinematics using finite-element simulation, supervised learning, and reinforcement learning.

Committee:

Nancy Pollard

Stelian Coros

Oliver Kroemer

Wen Sun