Grasping analysis for automated assembly based on a combination of gripping models and qualitative reasoning

Abstract

The work proposes a knowledge reasoning system for grasping analysis in terms of gripping models and qualitative reasoning. A matrix representation of object gripping attributes, aimed at establishing a correspondence to the gripper structure and performance, is developed. Use of possible decision-making based on incomplete information of object characteristics and gripper potential actions is included. The use of the concept of imitating human-based expert systems to model events and predict physical system actions with no precise numerical data is incorporated. Finally, a generic gripper developed to validate the proposed approach is discussed.