Grasp Synergy Analysis Based on Contact Area of Fingers Using Thermal Signatures

Abstract

We propose a new method for analyzing human grasping motion using an infrared camera. This technique simplifies the teaching of motion to robots based on the observation of human motion. In this method, the contact area on the object is extracted by observing the thermal signature captured by an infrared camera. To understand the intention of human behavior, we propose a grasping identification method using 3D thermal signatures. In addition, we perform a principal component analysis on the contact area and the center of gravity for the contact area of each finger. This method expresses the grasp motion space with a small number of parameters and can be used to enable easy correspondence between human and robot hands. We confirm experimentally that the proposed method is effective for teaching motion to robot.