Intelligent Conceptual Design of Robot Grippers for Assembly Tasks

D.T. Pham, N.S. Gourashi and E.E. Eldukhri

Manufacturing Engineering Centre, Cardiff University, Cardiff, UK

Presentation Outline

• Introduction
• Design Process
• Conceptual Design
• Intelligent Gripper Design System
• Example Application
• Conclusion

Introduction

• The paper presents a system that automatically configure optimum gripper systems for robotic assembly tasks
• It gives an overview of conceptual design and configuration problems
• It describles the system and its design procedure
• It gives an example application of the system to a simple end-effector design problem

Design Process

The design process consists of three main phases:

1. Problem formulation
2. Conceptual Design
3. Detailed Design

This paper focuses on the conceptual design process.

Conceptual Design

• Conceptual design is very complex and requires a very high-level of mental activity
• It is a decision making process where complete and precise information about the problem are not available
• It is a phase where engineering science, practical knowledge, production methods and commercial aspects need to be brought together
• It takes the goal and specifications of the design problem along with other constraints as inputs and gives in return a scheme or a concept as output

In this work, a case-based reasoning technique is used to reconfigure predifined robot gripper components into a final solution that satisfies the design problem.

Intelligent Gripper design System

The five design steps are: 

1. Importing CAD models into the intelligent design software
2. Recognising geometric structures of the imported models
3. Determining the orientations and all relevant dimensions of the compoments to be assembled
4. Grouping of similar components into families
5. Generating the single gripper capable of handling all components

an imported cylinder

Checkers on the Reference Box

Example Application

• The example illustrates the application of the design system to a simple gripper configuration problem
• The configuration problem is fully automated
• The user is only required to specify the components that need to be specified
• Four components of the same geometric class, material type, size family and range of weights are specified
• The system then proposes the optimum set of grippers for the given component
• Finally, the system generates a single gripper capable of handling all the specified components

Components to be handled

Generated Gripper

Conclusion and Further Work

• An approached for automated optimum design of gripping systems was described
• A five-step design procedure for fully automated conceptual design was explained
• An example application of the design system to a simple robot gripper was given
• The obtained gripper system is capable of handling more complex tasks
• Future work could focus on designing grippers using surface models of the components to be gripped