Neuro-Fuzzy Inference System based controller in 6 DoF in PUMA 560

Abstract

One of the problems in robotics field is, to find the orientation and framing of arm's end to desired location there for the new method must be find the solution of forward and inverse Kinematics (IK) problem. Traditional methods such as iterative, geometric and algebraic are used to solve (IK) inadequate if the joint structure of the robot is more complex. As the complexity of manipulator increases, solution of IK is difficult and computationally expensive. In this paper the use of the ability of Adaptive Neuro-Fuzzy Inference System (ANFIS) to learn from training data is implemented. Then the creation of ANFIS with limited mathematical representation of the system is performed. In the proposed method, forward kinematics relations of robot arm are used to obtain the data for training and testing of ANFIS. Simulation using MATLAB and microcontroller is used to show the effectiveness of the approach. Also the practical results which are carried out on PUMA 560 with 6DoF robot arm are used to show the effectiveness of the approach and proved that this method has good performance.