State and stiffness estimation using robotic fabrics

Abstract

Robotic fabrics are planar, fabric-based systems with embedded sensing and actuation functionalities. They are a useful, reconfigurable tool which can turn passive structures into active robots through surface-induced deformations. However, because of this flexibility, it is difficult to create empirical models for all possible configurations and host body materials that may be used with robotic fabrics. In this paper, we focus on the widely-applicable case of a continuum joint formed by wrapping a robotic fabric around a soft cylinder, and propose a model that is compatible with a variety of host body materials. The model is able take sensor data from the robotic fabric and then estimate both state and stiffness of the underlying structural material. We show the functionality of our model on three different materials: polyethelene foam, Dragonskin 10 Slow elastomer, and Smooth-Sil 935 elastomer. Simplified models that are able to provide both state and stiffness estimations are an important tool that can lead to advancements in control of soft robots.