Authors:

Richard Li

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Modern methods of finding stable grasp configurations for robot hands commonly use learning models to determine force closure contact point sets. To improve runtime for feasibility for real world use, some methods also incorporate analytical algorithms to filter and refine grasp configurations before processing in the learning model. However, most of the literature on these algorithms focus on parallel jaw grippers. These grippers have more straightforward algorithms, often using antipodal grasp criterion to take advantage of the two-contact point nature of the gripper. This study addresses the issue of finding and verifying force closure contact points for multiple fingered grippers by leveraging available information from object mesh and gripper models to quickly synthesize stable grasp configurations using an analytical algorithm. We use a signed distance format gripper representation combined with a point cloud representation of an object to optimize grasp configurations based upon force closure equation approximations. Previous papers indicate it is possible to find stable grasp configurations with object and gripper models within milliseconds on a modern desktop. We verify this using our variation of these equations with various object and gripper configurations. This research improves upon the feasibility for data driven robotic grasping of novel objects by optimizing the process of finding stable contact point sets to reduce runtime and complexity. Keywords: Robotic Grasping; Grasp Synthesis; Force Closure; Optimization; Smart Manufacturing

Source:

Purdue University / 2024

Topics:

No topics listed

Co-authors:

Richard Li