Determining the arm's motion angle using inverse kinematics models and adaptive neuro-fuzzy interface system

Abstract

Robotics technology is known as a great technology demand to be developed continuesly. One of the important things that need to be considered is the control of the motion of the robot. Movement predictions can be modeled in mathematical equations. Prediction based on learning logic is also very supportive of motion control systems, especially arm motion. In this study, the authors combined the two methods as the main study. The working principle of the arm is to take colored objects detected by the camera. In this study, we made arm four DOFs (Degree of Freedom), but only one DOF is controlled by ANFIS because the other three DOFs only move at two fixed angles. Two methods of determining the arm angle of motion used are inverse kinematics and ANFIS methods. The angle of motion and the position of the red object can be observed in real-time on the monitor with the interface in the MATLAB GUI. The angular output that appears in the MATLAB GUI is sent to Arduino in the form of characters, then, Arduino translates it into servo motion to the coordinates of the object detected by the camera. The results showed that the ANFIS method was more effective than the inverse kinematics model.