1: \begin{abstract}

2: %

3: Tracking Cartesian trajectories with end-effectors is a fundamental task in robot control.

4: %

5: For motion that is not known a priori, the solvers must find

6: fast solutions to the inverse kinematics (IK) problem for discretely sampled target poses.

7: On joint control level, however, the robot's actuators operate in a continuous domain,

8: requiring smooth transitions between individual states.

9: %

10: In this work we present a boost to the well-known Jacobian transpose method

11: to address this goal, using the mass matrix of a virtually conditioned

12: twin of the manipulator.

13: %

14: Results on the UR10 show superior convergence and quality of our dynamics-based solver against the plain Jacobian method.

15: Our algorithm is straightforward to implement as a controller, using present robotics libraries.

16: \end{abstract}

17: