Development of 3D elastic deformable object model for robot manipulation purposes

Abstract

This paper presents a mass-spring model applied in the manipulation of elastic deformable object for home service robot application. In this paper, we present a system used to fold a piece of rectangular cloth from a specific initial condition using robot. The cloth is modeled as a 3D object in a 2D quadrangular mesh based on a mass-spring system and its state is estimated using an explicit integration scheme that computes the particle position as a function of internal and external forces acting on the elastic deformable object. The state of the elastic deformable object under robot manipulation is currently tracked from the trajectory of the mass points in the massspring system model in a self developed simulator, which integrates a mass-spring model and a 5 DOF articulated robotic arm. To test the reliability of the model, the simulator is used to predict the best possible paths for the robotic arm to fold a rectangular cloth in two. In the test, the state of the object is derived from the model and then compared with practical experiment. Based on the test, the error is generally acceptable. Thus, this model can be used as an estimator for vision-based tracking on the state of an elastic deformable object for manipulation by home service robots. © (2012) Trans Tech Publications, Switzerland.