Sistemas computacionais inteligentes aplicados ao controle de um braço manipulador robótico simulado

Abstract

This work performs the control of a simulated autonomous robotic manipulator arm in the MatLab® software, both simulation and control are performed using intelligent computational systems. An Artificial Neural Network (Multilayer Perceptron) related to camera calibration is used. Genetic algorithm, Jacobian matrix and Denavit-Hartenberg method are used to program the direct and inverse kinematics of the manipulator arm and to simulate it. The union between the camera and the simulated robotic arm is performed in order to conclude the objective in a synergistic way. It has thus the ability to perform the detection of certain point by means of the camera, and achieve it using the robotic manipulator arm. The results presented are the errors obtained between the point reached and the desired point and the system response time. At the end of the paper, the proposed conclusions and future work are presented.