Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: how do you get around the wrong computed torque data in SimMechanics
Replies: 1   Last Post: Aug 27, 2014 6:30 PM

 Messages: [ Previous | Next ]
 jphoton Posts: 4 Registered: 8/22/14
how do you get around the wrong computed torque data in SimMechanics
Posted: Aug 22, 2014 11:47 PM

I am using Matlab/SimMechanics 1st generation.

-I use inverse dynamics to calculate the torques required to produce a specified motion for a model. Itâs a simple rotation of about 15 degrees.

-I then apply those torques to the same model in forward dynamics mode. However the rotation I obtain is about 345 degrees.

-According to this tutorial, this is actually normal.
http://www.mathworks.co.uk/help/physmod/sm/mech/ug/finding-forces-from-motions.html

So the questions:

1
In order to obtain the correct motion using the torques, do I then use only the torque data from the region of time that provides me with the correct motion and basically interpolate between the torques to obtain a smoother motion for the full time I require?? It seems like a good idea now but Iâm testing only one link of a robot, if I add all the links and a more complex motion I donât know if it will still make sense to do this.

2
The reason I need the torques is because I need to simulate the system in response to an external force. If I specify the motions in inverse dynamics mode (or even forward dynamics); it will simply give me back exactly the motion I require regardless of the external force. So I basically want to apply my motion (WITHOUT the external force) in inverse dynamics mode to obtain the torques. Then I apply the same torques to the model (WITH the external force) to see what happens to the motion.
Is this the right way to do it or is there another way??
And if it is indeed the right way, then this brings us back to question 1 above.

I would appreciate any help. Thanks in advance.
Joe.

Date Subject Author
8/22/14 jphoton
8/27/14 jphoton