Design optimization of Stewart platform for motion simulation systems using multi-objective genetic algorithm

Abstract

This paper utilizes the multi-objective genetic algorithm (MOGA) to optimize the design of Stewart platform. Particularly, we propose an optimization scheme based on two optimization objectives which consider the technical requirements of the parallel robot system when using in motion simulation systems. The first objective is the evaluation of workspace matching between the actual workspace and the desired workspace. The second objective evaluates the forces exerted on the legs of the manipulator in some special conditions of the load exerted on the platform. We proposed the MOGA based on nondominated sorting genetic algorithm II (NSGA-II) [1], but, here we include the effective mutation and crossover methods. As a result, when the realistic technical specifications of Stewart platforms in motion simulation systems is considered as inputs of MOGA, a set of optimal solutions called the Pareto front is attained. Finally, a schematic configuration is chosen to build a real Stewart platform.