Determination of inverse kinematic solutions for a 3 Degree of Freedom Parallel Manipulator using Machine Learning

Abstract

As the complexity of parallel manipulator increases, it becomes difficult to model it kinematically. Performing the inverse kinematic analysis using the algebraic approach for such manipulators are also a cumbersome task and nearly impossible for some cases. This paper presents the determination of the inverse kinematic solutions for a 3-PSS (P-Prismatic, S-Spherical) parallel manipulator using various Machine Learning approaches such as Multiple Linear Regression, Multi-Variate Polynomial Regression, Support Vector Regression, Decision Tree Regression and Random Forest Regression. The P has been underlined to indicate the active joint in the mechanism. The results are presented and compared with the experimental and analytical results to prove the efficiency of the proposed approach.