Quaternion Modeling of a Delta Planar Robot and Training of an Enhanced Multilayer Neural Network to Solve the Inverse Kinematic Problem

This paper presents the modeling of inverse kinematic problem related to motions a delta planar robot using algebra unitary Quaternions. The mathematical model resulting from analysis has an associated system 8 nonlinear algebraic equations with polynomial unknowns. Newton-Raphson method was used solve robot. Subsequently, robot, database constructed that relates Cartesian coordinates end effector angles and axes rotations links. train multilayer neural network in order have equivalent problem. A series experiments were performed obtain improved configuration by varying four training parameters. results obtained show trained can be studied