A Robotic Sensory System with High Spatiotemporal Resolution for Texture Recognition

Published in Nature Communications, 2023

Figure: A robotic sensory system mimicking the human sensory system for texture recognition. a) The biological sensory system of humans. b) The arti system of this study, for which the sensor can detect both static and dynamic pressures.

Abstract

Humans can gently slide a finger on the surface of an object and identify it by capturing both static pressure and high-frequency vibrations. Although modern robots integrated with flexible sensors can precisely detect pressure, shear force, and strain, they still perform insufficiently or require multi-sensors to respond to both static and high-frequency physical stimuli during the interaction. Here, we report a real-time artificial sensory system for high-accuracy texture recognition based on a single iontronic slip-sensor, and propose a criterion—spatiotemporal resolution, to corelate the sensing performance with recognition capability. The sensor can respond to both static and dynamic stimuli (0-400Hz) with a high spatial resolution of 15μm in spacing and 6μm in height, together with a high-frequency resolution of 0.02Hz at 400Hz, enabling high-precision discrimination of fine surface features. The sensory system integrated on a prosthetic fingertip can identify 20 different commercial textiles with a 100.0% accuracy at a fixed sliding rate and a 98.9% accuracy at random sliding rates. The sensory system is expected to help achieve subtle tactile sensation for robotics and prosthetics, and further be applied to haptic-based virtual reality and beyond.

Video

A portable and real-time sensory system for texture recognition 👇🏻

A portable and real-time sensory system for texture recognition using a prosthetic hand equipped with a slip-sensor 👇🏻

🔍 Fullscreen View | Download PDF from Google Drive 👆🏻


Recommended citation: Bai, N.*, Xue, Y.*, Chen, S., Shi, L., Shi, J., Zhang, Y., ... & Guo, C. F. (2023). A robotic sensory system with high spatiotemporal resolution for texture recognition. Nature Communications, 14(1), 7121.
Download Paper