Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : June 05, 2019 - June 08, 2019
Many studies on material estimation by a robot have been proposed, but appropriate movement on target object is required. On the other hand, when humans obtain material information, they can touch the object with appropriate contact force and scanning velocity according to the shape and mechanical properties of the object. In this paper, material estimation by using a wearable skin vibration sensor is proposed. The sensor allows users to touch an object by the bare fingertip. Ten kinds of materials were used in the experiment and estimation with convolution neural network was adopted. The correct answer rate of the learning model was about 60%, and it was seemed that the material information estimated with the sensor was relevant to the information on roughness and friction.
