Stiffness Control of Deformable Robots Using Finite Element Modeling

Margaret Koehler (1, 2) , Allison M Okamura (1) , Christian Duriez (2, 3)

1 Department of Mechanical Engineering [Stanford]
2 DEFROST - Deformable Robots Simulation Team
3 Inria@SiliconValley - Inria@SiliconValley

Due to the complexity of modeling deformable materials and infinite degrees of freedom, the rich background of rigid robot control has not been transferred to soft robots. Thus, most model-based control techniques developed for soft robots and soft haptic interfaces are specific to the particular device. In this paper, we develop a general method for stiffness control of soft robots suitable for arbitrary robot geometry and many types of actuation. Extending previous work that uses finite element modeling for position control, we determine the relationship between end-effector and actuator compliance, including the inherent device compliance, and use this to determine the appropriate controlled actuator stiffness for a desired stiffness of the end-effector. Such stiffness control, as the first component of impedance control, can be used to compensate for the natural stiffness of the deformable device and to control the robot's interaction with the environment or a user. We validate the stiffness projection on a deformable robot and include this stiffness projection in a haptic control loop to render a virtual fixture.

Mots clés: -Modeling; Compliance and Impedance Control; Haptics and Haptic Interfaces; Control; Learning for Soft Robots

Domaines: Robotique [cs.RO]

Margaret Koehler, Allison M Okamura, Christian Duriez. Stiffness Control of Deformable Robots Using Finite Element Modeling. IEEE Robotics and Automation Letters, 2019, 4 (2), pp.469-476. ⟨10.1109/LRA.2019.2890897⟩. ⟨hal-02079146⟩