On the calibration of deformation model of rheology object by a modified randomized algorithm

Abstract

There are many kinds of rheology objects in our living life. If such rheology objects are individually and completely modeled, we can flexibly deal with such various objects in a factory by a robotic manipulator controlling six degrees-of-freedom forces and moments and also operate them at a 3-D graphics world in a house by a human via a wonderful haptic device feeling the forces and moments. If such a system is developed after acquiring a rheology dynamic model, we can enjoy a clay work in a 3-D virtual environment. For this purpose, we calibrate deformation model of a rheology object by a modified randomized algorithm based on experimental data. The data is measured from a bread material pushed by a robotic manipulator exactly. The model is a 3-D voxel/lattice structure with many elements, and each element consists of two dampers and one spring. By changing three coefficients of dampers and spring, we can describe various material properties concerning to viscosity and elasticity. In this paper, by minimizing shape difference between real and virtual rheology objects in the near-optimal algorithm, we find a better set of coefficients. Using the calibrated model, we can feel three degrees-of-freedom forces and three degrees-of-freedom moments attracted from a calibrated virtual object in a haptic device and synchronously we can watch shape deformation of the object in a 3-D graphics animation.