Application of Soft Computing Techniques in the Design of Robot Grippers

The present paper presents a system used for performing design by making extended use of soft computing techniques. Throughout the paper the design of robotic grippers for flexible materials is used as a reference design case in both its conceptual and detailed design phases. The system contains three modules. The first module focuses on the conceptual design phase and is used for: a. the generation of a set of feasible solutions (concepts) of grippers for a set of pre-specified requirements which express the handling tasks that the gripper must accomplish and b. for the evaluation of the feasible solutions in order to find the best one according to the designers’ experience and some prespecified criteria. The evaluation process is based on a design index that includes several interacting criteria represented as relationships among fuzzy entities. The concept is then submitted to the second module called Case-DeSC (Case-based Design with Soft Computing). Case-DeSC is dedicated to the development of a detailed parametric design of the gripper concept by utilizing formal design case representation, fuzzy preferences on the design parameters and by searching for optimal solutions by deploying an evolutionary algorithm. The optimal solution passes over to a third module, capable of presenting a 3D visual representation of the parametric gripper concept via Virtual Reality Modelling Language (VRML). The proposed system has been developed in MATLAB.