An efficient data acquisition methodology for inverse dynamics model learning of manipulator based on analytical method. II

We consider a parametric physical model of a manipulator obtained from the rigid body dynamics using the analytical method. Our framework consists of the Denavit — Hartenberg method for the generation of manipulator workspace, the cubic polynomial method for the trajectory generation between two points, the Levenberg — Marquardt method for finding the required joint positions to reach the goal points and the Newton — Euler method for finding the required torque to execute the desired trajectory. The received datasets are validated by the results of simulation of kinematic and dynamic modeling of the tested manipulator.