Design and Optimization for a New Compliant Planar Spring of Upper Limb Assistive Device Using Hybrid Approach of RSM–FEM and MOGA

Abstract

This study develops a new multi-objective optimization design for a compliant planar spring that is used for upper limb assistive device for disabled people. The structure of spring is designed by arranging the leaf springs in a zigzag shape to permit a large deformation. The spring is a key element of the gravity-balanced mechanism (GBM) in the upper limb assistive device, exoskeleton and upper limb assistive robot. Static equilibration of the GBM is utilized to eliminate the working effort of the disable forearm. The spring is therefore used to adjust the stiffness and to maintain the equilibrium condition. To improve static performances, an effective hybrid approach of the response surface method, finite element method, and multi-objective genetic algorithm is developed. In comparison with common types of surrogates, the Kriging metamodel was found as a prefer black-box to approximate the objective functions. Subsequently, the multi-objective genetic algorithm is adopted to find out the Pareto-optimal set for the spring. The sensitivity of each design variables is then analyzed, and their significant contributions are conducted using the analysis of variance. The predicted results give a good agreement with the validation ones. It eliminates the modeling errors of the analytical method for synthesis and analysis of compliant mechanisms. The results facilitate an optimization design for the intelligent robotics systems for the assistive devices.