Deformation Planning for Robotic Soft Tissue Manipulation

This paper presents a model based approach to the soft tissue deformation planning. The deformable object is manipulated through boundary displacements induced by robot manipulators controlled in position. The manipulated boundaries are maneuvered such that the control points de-fined on the deformable object converge to the desired locations. The proposed control is based on a Jacobian trans-formation between the set of manipulated point displacements and the control point displacements computed using a meshless model (Reproducing Kernel Particle Method-RKPM) of the deformable object. RKPM is employed for this study as it has been proven to accurately handle large deformations and requires no re-meshing algorithms. Simulations show that a model with a coarse particle grid can produce Jacobian transforms that accurately control a more physically real and refined model. The next step is to perform a physical study on a tissue phantom interacting with a dual arm manipulator.