Neural network algorithm for workspace analysis of a parallel mechanism

Abstract

Purpose – This paper aims to combine and further develop different mathematical models of the workspace representation of 6 degrees of freedom parallel mechanisms and to bring a new point of view to existing workspace analysis methods through using neural networks (NN).Design/methodology/approach – For the orientation workspace of the 6‐3 SPM, discretization method is used which is based on Euler angles and the NN algorithm is applied.Findings – The workspace analysis is carried out in the direction perpendicular to the moving platform which is the most workable direction of 6‐3 Stewart platform mechanisms and NN algorithm has decreased processing time.Originality/value – The determination of the point, on that direction, at which the workspace is maximum, is outlined. It is the first time that the NN is used for classification of workspace of a parallel manipulator.