EM-Skin: An Artificial Robotic Skin Using Magnetic Inductance Tomography

摘要

Physical sensing by touch is essential for building intelligent artificial systems in robotic manipulation and human-robotic interaction. Inductive skins are being investigated as part of a major effort to develop the most robust and reliable touch sensors, primarily based on traditional inductive proximity sensing. Magnetic induction tomography (MIT) is an imaging system considered for medical diagnostics and industrial process monitoring. This article presents a novel electromagnetic-based skin (EM-skin) using the MIT imaging system. This is done by processing the mutual inductance data from a planar array sensing a skin-like medium, including an elastomeric medium that interfaces the MIT sensors with plates of metallic or magnetic touch elements. The paper demonstrates EM-based multi-touch, dynamical touch, and quantitative touch pressure sensing. MIT data is captured at 10 frames/s, so allowing for dynamical touch analysis. The EM-skin sensing area of 900 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> demonstrates a large area of sensing skin. The results show the successful reconstruction of dynamical sensing, where two applied cyclic touch points, with different frequencies are discriminately detected. Quantitively force sensing shows the detection of a minimum of 120 mN force, which translates to 0.38 kP of applied pressure in the described system. Further force calibration is carried out demonstrating the quantitative nature of the proposed EM skin. These results will open the way to a new generation of distributed and reliable soft skins that are versatile due to material design and processing.