Abstract

This paper presents a polymeric optical fiber-based tactile sensitive artificial skin designed for the replication of touch on a robotic finger surface. The artificial skin integrates a polymer optical fiber, inscribed with an array of four fiber Bragg gratings, inside a 2-mm thick silicone membrane to enable simultaneous detection of force magnitudes and contact localization around the sensing region. The presented skin was fabricated by utilizing the dimensions of a human index fingertip where a four-axis electronic platform was used to replicate the force exertion on the skin over an indenter of 2 mm in diameters. A machine learning algorithm, random forest regression, was employed to estimate force magnitude and contact location by analyzing the wavelength shift of the four FBGs. The proposed POF-based artificial skin is equipped with high-resolution force sensing capability, especially in confined areas while offering simple fabrication and good scalability for large area tactile sensing on curved and complex surfaces. These results demonstrate that the proposed skin is suitable for large area sensorization on robots to achieve dexterous object manipulations and explorations.