The Inverse Kinematics Evaluation of 6-DOF Robots in Cooperative Tasks Using Virtual Modeling Design and Artificial Intelligence Tools

Abstract

—This work aims at evaluating the inverse kinematics of a two-robot cooperative system using Matlab/SimMechanics-based simulations and artificial intelligence tools, namely the Levenberg-Marquardt (LM) optimization method. The artificial neural networks (ANN) thus constructed will replace the controllers of the six degrees of freedom (6-DOF) cooperative robots. Therefore, the entire cooperative system was designed in SolidWorks, taking into account all the dimensions necessary for kinematic modeling, then converted into Matlab/SimMechanics, and thanks to the manipulation of the model in this software, we will be able to extract the articulatory and operational data of the cooperative system in its workspace. The kinematic database of the robotic system is built in Matlab in order to train the ANN and implement it in Matlab/SimMechanics. Lastly, a test is performed in a collaborative task to evaluate the intelligent control error. The results obtained can be applied to and tested for the kinematic control of two real ABB IRB 120 cooperative robots.