Workspace identification using neural network for an optimal designed 2-DOF orientation parallel device

Abstract

The main purpose of the paper is to develop a mathematical model that generates the optimal geometrical parameters combination for a 2-DOF parallel mechanism, and to perform a solution to generate the workspace by using neural networks as a performant alternative to the workspace representation based on inverse kinematic model. The paper describes three algorithms that lead to the final solution and an initial testing was made on a functional model of parallel mechanism. © Springer Science+Business Media Dordrecht 2013.