Graduation Year

2024

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Computer Science and Engineering

Major Professor

Yu Sun, Ph.D.

Committee Member

Yasin Yilmaz, Ph.D.

Committee Member

Dmitry Goldgof, Ph.D.

Committee Member

Paul Rosen, Ph.D.

Committee Member

Zhao Han, Ph.D.

Committee Member

Joe Askren, Ph.D.

Keywords

Dexterous Manipulation, Motion and Path Planning, Logistics, Multi-fingered Hands

Abstract

As robots become increasingly integrated into real-world applications such as warehousing, fulfillment centers, and manufacturing, the need for efficient and adaptable robotic systems grows. One of the key challenges is enabling robots to grasp multiple objects simultaneously, as this significantly boosts the efficiency of tasks like batch picking, sorting, and object transferring, reducing both time and energy consumption. This dissertation presents a comprehensive multi-object grasping (MOG) pipeline that includes pre-grasp selection, end-pose selection, grasping synergy calculation, and a data-driven model for estimating the number of objects being grasped. Central to this work is the development of the Experience Forest structure, which organizes finger movement sequences into multiple trees, enhancing the efficiency and reliability of MOG by propagating success or failure results through the grasping process. Additionally, the research introduces a novel approach to object transferring, modeling the problem as a Markov decision process (MDP) to define specific grasping actions for efficiently transferring quantities larger than the capability of a single grasp. Furthermore, based on insights from both human demonstrations and robotic MOG solutions, a taxonomy of 12 MOG types is proposed, classified into shape-based and function-based groups. Finally, we introduce another MOG strategy to grasp objects from the top surfaces of the object pile, such that the system can deal with more types of objects and more complex environment. Experimental evaluations using a UR5e robotic arm and Barrett hand in both simulated and real environments demonstrate a significant $60\%$ increase in efficiency for object transfer tasks. These advancements offer a robust framework for improving the adaptability and efficiency of robotic systems in unstructured environments.

Download

Share

COinS