Towards Reconfigurable and Adaptive Soft Robots via Hybrid Materials, Designs and Mechanisms
Inspired by biological systems, soft robotics has become a new research field that builds compliance and conformability into soft machines and devices. Soft robots are capable of functions that traditional rigid robots cannot achieve. The implementation of intrinsically soft materials allows for deformable robot bodies and adaptable robot kinematics, enabling applications such as flexible sensing technologies, robust grasping systems, and safe human-machine interactions. However, soft robots also face challenges associated with their compliance: lack of precision in kinematics, lack of reconfigurability, and lack of structural strength can all hinder soft robot capabilities. While significant research has focused on the control of actively tunable material properties and soft actuation methods for soft robots, in this work we focus on the design of soft reconfigurable robot architectures that utilize highly manipulatable soft materials and rigid components into hybrid soft/rigid building blocks. In this dissertation, the author provides new design paradigms for future reconfigurable and adaptive soft robots via the development of hybrid materials, designs, and mechanisms. This topic is addressed through three main design principles: a) Geometric reconfiguration of two-dimensional rigid and soft materials leading to variable stiffness change, and multifunctional robot laminates. b) Additive manufacturing combined with flexible laminates to achieve hybrid three-dimensional rigid and soft structures for on-demand stiffening and adaptive robot grasping and locomotion. c) Pinching of soft tubular structures enforced by rigid constraints to achieve on demand formation of reconfigurable virtual joints. Through these studies, the author seeks to present novel reconfigurable mechanisms that utilize combinations of passively compliant and rigid materials and simple activation approaches exploiting these materials. These same design paradigms can be extended to nonlinear programmable material properties enhancing robot performances in the future.