Simulink Simulation Environment of the Franka Emika Panda Robot

When developing custom controllers for robots, it is helpful to simulate and test them before implementing them on real hardware. The Franka Emika Panda, a robotic research platform, provides a Matlab-Simulink interface for programming the robot using the Franka Matlab library. However, the Franka Matlab library only supports the building and execution of control algorithms in external mode on real hardware and does not support offline testing and tuning of controllers beforehand. Thus, this paper proposes a Simulink model based on the Matlab Robotics System Toolbox that allows the development and tuning of controllers for the Franka Emika Panda robot before they are ported to the real hardware. The model used in simulation mimicks the kinematic, but also dynamic behavior of the robot to best imitate the actual robot. To illustrate the usage of the simulation environment, two basic motion controllers and one more advanced energy-shaping controller are implemented and tested first in simulation and then on the real platform. Obtained results show high similarities between simulation and experiments and demonstrate the usefulness of the model in supporting the design of controllers.