Flexible electronic brush: Real-time multimodal sensing powered by reservoir computing through whisker dynamics

Abstract

Multimodal sensing using soft body dynamics plays a crucial role in controlling soft robotic motions. An intriguing application of such soft robot control is to mimic whiskers and digitize soft body motion through whisker dynamics. The challenge herein is to simultaneously monitor the directions, speed, force, and slip information of the whisker motion. The existing whisker-like sensors cannot detect slip information effectively. To address this challenge, this study develops a multitasking electronic brush (e-brush) composed of bundle of whiskers powered by reservoir computing (RC). Four pressure sensors are integrated into the brush to monitor its motion, speed, force, slip, and target surface. These sensors can provide long-term, low-pressure detection as low as 50 pascals, allowing for the precise monitoring of brush movements. A RC algorithm is developed to extract multiple brush motion parameters, including the slip. As a proof of concept for multitasking e-brush, the motion trajectory of handwriting is successfully detected.