Forward Kinematics of a 6x6 UPU Parallel Mechanism by ANFIS Method

Abstract

In this paper, we address the forward kinematics solution of 6DOF 6x6 UPU parallel mechanism using three different approaches. In contrast to the case of serial robots, the forward kinematics of parallel manipulators contain highly coupled nonlinear equations which are possible to solve analytically only in very special cases. In the proposed paper, Adaptive-Network-based Fuzzy Inference System (ANFIS), Multilayer Perceptron (MLP) and Radial Base Function (RBF) Neural Network methods are implemented to reach the forward kinematics solution of a low payload flight simulator. Solution accuracies have been compared, and it has been concluded that ANFIS shows better performance compared to MLP and RBF networks.