forces in joints

Uwe Mittag · Post by **Uwe Mittag** » Wed Jul 03, 2013 4:00 am

Thanks Ayman and Jiang,

again shortly what I want to implement:
two sticks with a principally free joint in between, but the joint producing resistance when bended (bushing force?). And of course I need the possibility to induce this bending by pressing against the body who is not ground.
Below my approach so far, using an actuator. Actuator does not seem to be the right choice as Ayman said, but how else to apply a simple force?

Thanks in advance and best kind regards
Uwe

Here the model (again my question, if there is no way to upload a textfile for this purpose). As far as I believe, it has two cylinders scaled to 2cm length with a rotational stiffness of 0.1 (whatever that means). I have already reduced that value, because I was afraid not to see the effect because the stiffness is to high.

<?xml version="1.0" encoding="UTF-8" ?>
<OpenSimDocument Version="30000">
<Model name="multisegment_test">
<credits>The OpenSim Development Team (Reinbolt, J; Seth, A; Habib, A; Hamner, S)</credits>
<publications>This is an example model distributed with OpenSim</publications>
<length_units>meters</length_units>
<force_units>N</force_units>

<gravity> 0 0 0</gravity>

<BodySet>
<objects>
<Body name="ground">
<mass>0</mass>
<mass_center> 0 0 0</mass_center>
<inertia_xx>0.1</inertia_xx>
<inertia_yy>0.1</inertia_yy>
<inertia_zz>0.02</inertia_zz>
<inertia_xy>0</inertia_xy>
<inertia_xz>0</inertia_xz>
<inertia_yz>0</inertia_yz>

<Joint />
<VisibleObject>

<GeometrySet>
<objects>
<DisplayGeometry>

<geometry_file>cyl_small.stl</geometry_file>

<color> 1 1 1</color>

<texture_file />

<transform> -0 0 -0 0 0 0</transform>

<scale_factors> 0.01 0.01 0.01</scale_factors>

<display_preference>4</display_preference>

<opacity>1</opacity>
</DisplayGeometry>
</objects>
<groups />
</GeometrySet>

<scale_factors> 1 1 1</scale_factors>

<transform> -0 0 -0 0 0 0</transform>

<show_axes>false</show_axes>

<display_preference>4</display_preference>
</VisibleObject>
</Body>
<Body name="cyl2">
<mass>0.1</mass>
<mass_center> 0 0 0</mass_center>
<inertia_xx>0.1</inertia_xx>
<inertia_yy>0.1</inertia_yy>
<inertia_zz>0.02</inertia_zz>
<inertia_xy>0</inertia_xy>
<inertia_xz>0</inertia_xz>
<inertia_yz>0</inertia_yz>

<Joint>
<FreeJoint name="joint1">

<parent_body>ground</parent_body>

<location_in_parent>0 0 0.01</location_in_parent>

<orientation_in_parent>0 0 0</orientation_in_parent>

<location>0 0 -0.01</location>

<orientation>0 0 0</orientation>

<CoordinateSet>
<objects>
<Coordinate name="_xRotation">

<motion_type>rotational</motion_type>

<prescribed>false</prescribed>
</Coordinate>
<Coordinate name="_yRotation">

<motion_type>rotational</motion_type>

<prescribed>false</prescribed>
</Coordinate>
<Coordinate name="_zRotation">

<motion_type>rotational</motion_type>

<prescribed>false</prescribed>
</Coordinate>
<Coordinate name="_xTranslation">

<motion_type>translational</motion_type>

<prescribed>false</prescribed>
</Coordinate>
<Coordinate name="_yTranslation">

<motion_type>translational</motion_type>

<prescribed>false</prescribed>
</Coordinate>
<Coordinate name="_zTranslation">

<motion_type>translational</motion_type>

<prescribed>false</prescribed>
</Coordinate>
</objects>
<groups />
</CoordinateSet>

<reverse>false</reverse>
</FreeJoint>
</Joint>
<VisibleObject>

<GeometrySet>
<objects>
<DisplayGeometry>

<geometry_file>cyl_small.stl</geometry_file>

<color> 1 1 1</color>

<texture_file />

<transform> -0 0 -0 0 0 0</transform>

<scale_factors> 0.01 0.01 0.01</scale_factors>

<display_preference>4</display_preference>

<opacity>1</opacity>
</DisplayGeometry>
</objects>
<groups />
</GeometrySet>

<scale_factors> 1 1 1</scale_factors>

<transform> -0 0 -0 0 0 0</transform>

<show_axes>false</show_axes>

<display_preference>4</display_preference>
</VisibleObject>
</Body>
</objects>
</BodySet>

<ConstraintSet>
<objects />
<groups />
</ConstraintSet>

<ForceSet>
<objects>
<BushingForce name="Joint1Bushing">

<body_1>ground</body_1>

<body_2>cyl2</body_2>

<location_body_1>0 0 0.01</location_body_1>

<orientation_body_1>0 0 0</orientation_body_1>

<location_body_2>0 0 -0.01</location_body_2>

<orientation_body_2>0 0 0</orientation_body_2>

<rotational_stiffness>0.1 0.1 0.1</rotational_stiffness>

<translational_stiffness>0 0 0</translational_stiffness>

<rotational_damping>0 0 0</rotational_damping>

<translational_damping>0 0 0</translational_damping>
</BushingForce>
<PointActuator name="bending_force">

<body>cyl2</body>

<point>0 0 0.01</point>

<point_is_global>false</point_is_global>

<direction>1 0 0</direction>

<force_is_global>true</force_is_global>

<optimal_force>80</optimal_force>
</PointActuator>
</objects>
<groups />
</ForceSet>

<MarkerSet>
<objects />
<groups />
</MarkerSet>

<ContactGeometrySet>
<objects />
<groups />
</ContactGeometrySet>
</Model>
</OpenSimDocument>

Jiang Ping · Post by **Jiang Ping** » Wed Jul 03, 2013 8:55 am

Hi,

You specified actuator but forgot to specify a controller to control it.
If you just want to add a force to the body, you can try this:

<PrescribedForce>

<isDisabled>false</isDisabled>

<body></body>

<pointIsGlobal>false</pointIsGlobal>

<forceIsGlobal>true</forceIsGlobal>

<FunctionSet name="forceFunctions">
<objects />
<groups />
</FunctionSet>

<FunctionSet name="pointFunctions">
<objects />
<groups />
</FunctionSet>

<FunctionSet name="torqueFunctions">
<objects />
<groups />
</FunctionSet>
</PrescribedForce>

best regards,
Jiang

Uwe Mittag · Post by **Uwe Mittag** » Thu Jul 04, 2013 4:21 am

Thank you Jiang,

it seems to work now. I first had problems to find how to manage the functions for the prescribedforce, but I found a functionset somewhere in the examples. But I am astonished, how long the integration is running for this little task. It is at 100% since a long time and always working. But the behaviour is as expected: moving to a bended situation and staying at an equilibrium position.

Thanks and kind regards
Uwe

Fabian Bayerlein · Post by **Fabian Bayerlein** » Thu Jul 04, 2013 1:26 pm

Hallo Uwe,

see picture for uploading of attachments.

Best,
Fabian

Jiang Ping · Post by **Jiang Ping** » Wed Jul 10, 2013 4:11 pm

mittag wrote:Thank you Jiang,

it seems to work now. I first had problems to find how to manage the functions for the prescribedforce, but I found a functionset somewhere in the examples. But I am astonished, how long the integration is running for this little task. It is at 100% since a long time and always working. But the behaviour is as expected: moving to a bended situation and staying at an equilibrium position.

Thanks and kind regards
Uwe

Hi,

I'm not familiar with the .xml.
If you want to add a constant force on a body, try this

PrescribedForce *f=new PrescribedForce(*body yourbody)
f->setForceFunction=(new constant(fx),new constant(fy),new constant(fz))
f->setPointFunction=(new constant(px),new constant(py),new constant(pz))
model.addforce(f)

Hope this will help you, thanks

Jiang

Uwe Mittag · Post by **Uwe Mittag** » Tue Jul 23, 2013 12:29 am

Hi thanks,

I have a 6-segment tibia now working more or less (using a prescribed force), but the dynamics runs seems to crash after some few ten of milliseconds. Up to may be 20 millisecs it reacts on stop commands and I can see the expected behaviour of bending with small oscillations. Would be nice to find the reason for that.

Now I want to integrate my bending tibia into one of your gate models. Is it possible to define the 6-segment tibia together with its bushing forces and joint as a body group, that can be aligned and scaled as a whole?

Thanks again and best regards
Uwe

Ayman Habib · Post by **Ayman Habib** » Thu Jul 25, 2013 1:21 pm

Hi Uwe,

If you're using our components (so we can reproduce the crash you reported in house), then please file a bug report and attach the model and all necessary files and we'll troubleshoot.

As it stands, the scale tool works on individual segments rather than groups although you can apply the same scale factor to multiple segments. BushingForces do not scale altogether so you'll have to scale their geometry and properties manually.

Hope this helps,
-Ayman

Uwe Mittag · Post by **Uwe Mittag** » Wed Nov 12, 2014 8:03 am

Hi,
I found the time to continue with this subject. I have added my segments manually to the model. It looks nice when I load the model. But when I run the inverse kinematics, the joints between my segments do not behave as they should. They show lot of translation in the joints. The bushing forces do not seem to work in the inverse kinematics. How can I influence the translation stiffness of my free joints in the inverse kinematics?

Best regards
Uwe

Ajay Seth · Post by **Ajay Seth** » Wed Nov 12, 2014 11:49 am

An inverse kinematics analysis is a purely kinematic analysis. There is no evaluation of any forces (no system dynamics). Specifically, it is the solution for the generalized (or internal) coordinates (joint angles, etc...) that match measurements of spatial (or absolute) coordinate measurements such as marker positions in ground (lab). If you are looking to have a dynamical system (model dynamics) follow experimental data then that it is generally the solution to a tracking problem.

CMC is a tracking controller, but it tracks the internal coordinates (not markers) from IK. It also has the requirement that each coordinate be controllable- that is each coordinate is accelerated by at least one actuator. That is why residual actuators are applied to the pelvis dofs.

Uwe Mittag · Post by **Uwe Mittag** » Thu Nov 13, 2014 4:51 am

Thank you Ajay,

I have now introduced very tight ranges for the translation DOF of my tibia segments and I checked carefully, whether parent and child positions of the joints are at the right place using a marker. Narrow translation ranges should have an effect already for the inverse kinematic, is'nt it? But again the segments have been shifted much more than the joints should allow. Do you have an example for me how to configure freejoints properly?

Kind regards
Uwe

forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints

Re: forces in joints