Trunk tutorial：How to add a translational motion in InverseMode？

[TeoREN0217]

Hi！
I‘m working on a soft arm based on Trunk tutorial and I want to add a translational motion on the Trunk root in the Inverse Mode. I have found the SlidingActuator but didn't found some examples to use it. Will the QPInvProblemSolver work on both cable and sliding actuators at the same time? How could I add the slidingActuators at the root position?

[EulalieCoevoet]

Hi @TeoREN0217,

Yes you can use multiple actuation types at the same time. That's no problem.

For the translational motion, I suggest you to rigidify the root of the robot first. You can have a look at this tool from STLIB to help you do that. Then I would add the SlidingActuator to the corresponding rigid frame.

You're right about the missing example. I've just added this one, which will be available in the next release:

def createScene(rootnode):
    from splib3.animation import animate, AnimationManager
    rootnode.addObject(AnimationManager(rootnode))
    rootnode.addObject('RequiredPlugin', pluginName=['SoftRobots', 'SoftRobots.Inverse', 'SofaPython3'])
    rootnode.gravity = [0, 0, 0]
    rootnode.addObject('DefaultVisualManagerLoop')
    rootnode.addObject('FreeMotionAnimationLoop')
    # Add the inverse solver
    rootnode.addObject('QPInverseProblemSolver')

    cube = rootnode.addChild('Cube')
    cube.addObject('EulerImplicitSolver')
    cube.addObject('SparseLDLSolver')
    cube.addObject('GenericConstraintCorrection')
    cube.addObject('MechanicalObject', template='Rigid3', position=[-10, 0, 0, 0, 0, 0, 1])
    cube.addObject('UniformMass', totalMass=0.01)
    # SlidingActuator needs a PartialFixedConstraint to fix the other directions
    cube.addObject('PartialFixedConstraint', template='Rigid3', fixedDirections=[0, 1, 1, 1, 1, 1])
    cube.addObject('SlidingActuator', template='Rigid3', indices=0, direction=[1, 0, 0, 0, 0, 0], maxDispVariation=0.1)

    visual = cube.addChild('Visu')
    visual.addObject('MeshOBJLoader', filename='mesh/cube.obj', scale=10)
    visual.addObject('OglModel', src='@MeshOBJLoader')
    visual.addObject('RigidMapping')

    # Set an effector for the inverse resolution
    effector = cube.addChild('Effector')
    effector.addObject('MechanicalObject', position=[10, 0, 0],  # front of the cube
                       showObject=True, drawMode=2, showObjectScale=1)
    effector.addObject('PositionEffector', indices=0, effectorGoal=[0, 0, 0])
    effector.addObject('RigidMapping')

    # Set a target for the effector that can be moved interactively
    target = rootnode.addChild('Target')
    target.addObject('EulerImplicitSolver', firstOrder=True)
    target.addObject('CGLinearSolver', tolerance=1e-5, iterations=20, threshold=1e-5)
    target.addObject('MechanicalObject', position=[30, 0, 0], showObject=True, drawMode=2, showObjectScale=3)
    target.addObject('PartialFixedConstraint', template='Rigid3', fixedDirections=[0, 1, 1, 1, 1, 1])
    target.addObject('UniformMass', totalMass=0.1)
    target.addObject('UncoupledConstraintCorrection')

    # Link the effector goal to the target
    effector.PositionEffector.effectorGoal.setParent(target.MechanicalObject.position)

[TeoREN0217]

Hi~Thanks for your reply! According to that I'm trying rigid my robot root and I extended the root like 50mm. The original root was located in [0,0,0] and the extended part is in the nagative parts of Z-axis. I'm confused about two params in the rigidify function which was the frames and groupIndices. How could I use [[rx,ry,rz], [qx,qy,qz,w]] or [[x,y,z,rx,ry,rz], [x,y,z,qx,qy,qz,w]] to represent the extended part? And how to choose the groupIndices?

[EulalieCoevoet]

For the groupIndices you can use a BoxROI for instance. It will help you select the points you want to rigidify.
For the frame, you can let it compute the geometric center of the points, and maybe just give an orientation aligned with the frame of reference. So [0,0,0,1] for a quaternion or [0,0,0] for Euler angles.

"""
            ...
            :param list frames: array of frames. The length of the array should be equal to the number
                                of rigid component. The orientation are given in eulerAngles (in degree) by passing
                                three values or using a quaternion by passing four values.
                                [[rx,ry,rz], [qx,qy,qz,w]]
                                User can also specify the position of the frame by passing six values (position and orientation in degree)
                                or seven values (position and quaternion).
                                [[x,y,z,rx,ry,rz], [x,y,z,qx,qy,qz,w]]
                                If the position is not specified, the position of the rigids will be the barycenter of the region to rigidify.
"""

[TeoREN0217]

Or can I use the fixExtermity() part (such as extending the box) instead of the rigid method?