Silicone Ethernet (SEth): a Nervous System for Robotic Touch

TL;DR

Silicone Ethernet (SEth) introduces a wireless, conductive silicone-based neural network for robotic touch sensing that enables energy harvesting, data communication, and touch detection within soft or hard robot skins.

Contribution

The paper presents a novel wireless neural network embedded in conductive silicone, enabling integrated sensing, communication, and energy harvesting for robotic touch applications.

Findings

Achieves data rates of 100 kbps with sub-μW receive power

Supports capacitive touch and presence sensing inherently

Provides energy harvesting for multiple message transmissions at 1 meter range

Abstract

Fine-grained robotic touch sensing is essential for tasks such as robot-human interaction and the handling of hazardous materials. Yet, the sense of touch of robots is limited by the cost and complexity of routing cables to embedded sensors. This paper tackles this problem by contributing Silicone Ethernet (SEth), a wireless solution for touch sensing, communication and power transfer within a conductive silicone substrate. SEth~\emph{neurons} require no battery and deliver computation, communication, sensing and energy harvesting within a compact package. These neurons are placed into an undifferentiated conductive silicone substrate which may form the entire body of a soft robot, or an outer `skin' for hard robots. Our evaluation shows that SEth achieves data rates of 100\,kbps with sub-$\mu$W receive and mW-scale transmit power. Exploiting the unique properties of the conductive silicone substrate, SEth provides prioritized traffic arbitration similar to that found in wired control networks such as CAN. The SEth network inherently supports capacitive touch and presence sensing and neurons can harvest sufficient energy to transmit 10s of messages per second at a range of 1\,m. Considered in sum, these features open new degrees of freedom in touch sensing for soft robots.