Neuro Fuzzy Concepts and Design (Case Study)

Abstract

Abstract This research presents a comparison between desired and actual trajectory used to adaptive control to robot manipulators which can be used in soldering square shape. In this thesis, the path trajectory obtained using the Adaptive Neuro Fuzzy Inference System (ANF IS) is called the actual path trajectory. ANFIS has two essential parameters; the first parameter is the number of fuzzy set, and the second parameter is the number of training. The training data of manipulator used in ANFIS is calculated by forward kinematics for two links worked in two dimensions. The desired path trajectory of manipulator is obtained by inverse kinematics method. Many comparisons can be made between desired and actual path trajectory. The execution of comparisons requires various values for both number of membership function and number of training. Finally, the suitable path trajectory selected by comparison between coirelation coefficients for both desired and actual path trajectory, the difference between desired and actual path trajectory is observed. And all the data used in comparisons are computed by MATLAB and Excel. Results obtained are of good value at number of training equal 75, and number of fuzzy set equal 9.