ANFIS based kinematic analysis of a 4-DOFs SCARA robot

Abstract

Robot kinematics plays a crucial role in recent advancements within industrial sectors and numerous medical applications. Finding forward kinematics, using DH convention is an easy task. As compared to forward kinematics, finding the inverse kinematics solution is far more challenging problem, especially when degrees-of-freedom (DOFs) are more. That is why; there is no general solution to the inverse kinematic problem of a given serial manipulator. This led to the development of alternate technique like fuzzy inference system (FIS) and neural network approach (NNA). This paper uses the combination of above two techniques, called as adaptive neuro fuzzy inference system (ANFIS), along with Gaussian membership function, in order to address the kinematic analysis of a 4-DOFs SCARA robot. The inverse kinematic solutions obtained using ANFIS are further utilized for desired path generation by the SCARA robot. Further, the complete analytical solution is developed in MATLAB environment for the validation purpose. It has been demonstrated with simulation runs that ANFIS results are satisfactory and are found in close approximation with analytical solutions.