A simulation-based method using artificial neural networks for solving the inverse kinematic problem of articulated robots:

摘要: In our article, inverse kinematic problem of a plasma cutting robot with three degree freedom is solved using artificial neural networks. Artificial network was trained joint angle values according to cartesian coordinates (x, y, z) end point robotic arm. The Levenberg–Marquardt training algorithm applied educate network. To validate the designed network, it tested new test data set which not in training. A simulation performed on three-dimensional model MSC.ADAMS software obtained from test. It revealed this that trajectory torch agreed well desired trajectory.