This dissertation propose a new fluidics based framework to determine a continuum between preshaping and grasping so as to appropriately preshape a multi-fingered robot hand for creating an optimal initialization of grasp. The continuum of a hand preshape closing upon an object that creates an initial object motion tendency of the object based on the impact moment patterns generated from the fingers is presented. These motion tendencies should then be suitable for the proper initiation of the grasping task. The aim is motivated by human like behavior where we preshape and land on an object to initiate a certain grasping behavior without losing the continuum during the " / preshaping to grasping" / phases.
The continuity of momentum transfer phenomena is inspired by fluid dynamics that deals with fluid flow. We have adapted governing equations based on the physical principles of the fluid flow to generate momentum transfer from the robotic fingers, closing upon the object surface to fluid medium particles then from these fluid medium particles to the grasping object. Smoothed Particular Hydrodynamics (SPH) which is a mesh free particle method and finite volume approximation is used to analyze fluid flow equations. The fingers of the robotic hand and object are modeled by solidified fluid elements and also can be compliance.
For evaluating the optimal grasp initialization of different hand preshape, we propose a decision support system consisting of artificial feed forward neural network based on the moment distribution on the object determines either : 1) given initial position and orientation of a robot hand, what preshape is suitable for generating a desired moment distribution on the surface of a given object in order to trigger a desired rotation in a desired direction when approaching with this preshaped hand or 2) given a predetermined hand preshape what initial position, orientation and hand aperture are suitable to generate a desired rotation upon approach and without causing the retroceeding of the object.
| Identifer | oai:union.ndltd.org:METU/oai:etd.lib.metu.edu.tr:http://etd.lib.metu.edu.tr/upload/12615138/index.pdf |
| Date | 01 October 2012 |
| Creators | Ozyer, Baris |
| Contributors | Erkmen, Ismet |
| Publisher | METU |
| Source Sets | Middle East Technical Univ. |
| Language | English |
| Detected Language | English |
| Type | Ph.D. Thesis |
| Format | text/pdf |
| Rights | To liberate the content for METU campus |
Page generated in 0.0022 seconds