Electronic Skin as Human–Robot Interface for Human Grasp Recognition in Home-Care Robots: Introducing a Complete Set of Resources

Abstract

Population aging has highlighted the significance of developing home-care robots to assist human life. However, traditional home-care robots work according to fixed procedures, which lack human–robot interaction to mimic human hand action. Human–robot interfaces provide an opportunity for interactions between robots and users, which play an important role in home-care robots. Traditional human–robot interfaces are based on rigid, cumbersome, and high-priced machines, which prevent their wide application in performing home-care tasks. An e-skin sensor with 11 stretchable resistors evenly distributed on the back of the human hand, which could be applied to monitor human finger motion and control robotic fingers to dexterously grasp diverse objects, is presented in this article. A 3D printer is employed to pattern high-performance wires into structurally soft and stretchable filamentary serpentines. The stretchable resistors are fabricated by a thin sensing film of mixed multiwalled carbon nanotubes (MWCNTs) and silicone elastomer. The e-skin sensor is applied in a hand-in-the-loop system for controlling a robotic hand to help grasp an object synchronously. A robot system is developed to connect the e-skin with the robotic hand to perform grasp actions with a success rate of more than 80%. By analyzing the experimental results of the e-skin sensor and robotic hand grasp, this work obtains insights showing that the skin sensor can help engineers design a home-care assistant plan for patients according to monitoring value or a human–robot interaction mode by robotic fingers mimicking human finger actions.