Four Universities and two industrial partners are involved in this project (U.W.C., U.R., T.U.B., U.P., Fazan, Image).

Development of an Expert System in designing of grippers for handling non-rigid materials.

The main aim of this research is to provide users of robots with a systematic approach for the design of grippers for handling non-rigid materials. The expert system will be implemented in a knowledge-based system called HOMER.

The objectives of the research are:

• To address specific non-rigid materials handling tasks which are difficult to automate and propose engineering solutions. The tasks will be specified by industrial partners.
• To develop an expert system for the design of grippers for handling non-rigid materials.
• To develop new grasping techniques for non-rigid materials.
• To improve current gripper design techniques specifically for the handling of non-rigid materials.
• To verify the developed knowledge-based system at the industrial partners’ site to prove its applicability and adaptability in real-world manufacturing environments

The Robotics Group of the University of Patras is interested in the following three tasks:

• Development of a Knowledge-based System for the generation of assembly strategies.
• Simulation of robot cells for assembly of a valve produced by Zita Ltd.
• Development of a Hybrid Force/Position Controller for the implementation of assembly strategies.

Knowledge Based System

• The knowledge-based system accepts as input IGES files where the parts are containing geometrical data for the parts to be assembled.
• The parts are classified according to their Geometric Features.
• The parts for each assembly task are classified as either moving or stationary.
• The C-Frames is defined on the moving part.
• The artificial constraints of each assembly subtask are formulated according to the contact case between the moving and the stationary part.

A special algorithm for the identification of the contact case is developed. This algorithm inputs the signal from the force sensor. In addition, a fuzzy decision system has been developed in order to reason about the feasibility of the assembly with respect to the tolerances between the two parts and the robot repeatability error.