Oscillation of robot with masses

Typically: "How do I... ", "How can I... " questions
Post Reply
Schindler
Posts: 27
Joined: 16 Jul 2013, 09:59

Oscillation of robot with masses

Post by Schindler »

I have build a robot with inverse kinematics. As long as I do not put any weight on it, the movement is rather straight and I might see only a few vibrations of the robot's head.
If I grip or assemble a little weight, the robot starts oscillating heavily.
I am not able to solve the problem by changing the simulation mode or the "dt".

In the examples and in the robots which are delivered with the software I saw that these robots are not build of grouped pure shapes but they are made of merged non-pure shapes which are static. Furthermore the hybrid mode of the inverse kinematics is switched off.
For my robot I used only grouped pure shapes, I selected the hybrid mode for inverse kinematics, the motors are enabled. These shapes weigh from about 3 to 7. The motor's forces are set from 400 to 1500.

Is there any way to find the optimal constellation of moment/force per each motor joint to the shapes that connect joints so that the robot does not oscillate?

Thanks for your support

coppelia
Site Admin
Posts: 10339
Joined: 14 Dec 2012, 00:25

Re: Oscillation of robot with masses

Post by coppelia »

Hello,

the static non-pure shapes you are referring to are not involved in the dynamics simulation: they are just the visual part, and the part that is used for distance calculation, etc. But they are dynamically not involved.

About your problem, it is difficult to say what is going on. I guess you followed all the dynamics design considerations?

If possible, post your scene via google drive or dropBox.

Cheers

Schindler
Posts: 27
Joined: 16 Jul 2013, 09:59

Re: Oscillation of robot with masses

Post by Schindler »

Hello,

I took the dynamic design advices into consideration.
Is it helpful to manipulate the moments of inertia manually?

You can find my robot here:
https://dl.dropboxusercontent.com/u/469 ... tstest.ttt

coppelia
Site Admin
Posts: 10339
Joined: 14 Dec 2012, 00:25

Re: Oscillation of robot with masses

Post by coppelia »

Hello,

try enabling the display of contact points in the dynamics dialog. You will get an idea of the main problem: most adjacent shapes are colliding. To avoid this, use a different local respondable mask for adjacent shapes (in the shape dynamics dialog).

Then, you are using the IK functionality for large movements, and the IK functionality will not know which path to take to avoid joint limits for instance. What you rather need here is a motion planning calculation. This functionality is in preparation and should be available in next release. In the mean time you will have to figure out the configurations of your robot for the 2 tip positions manually (or iteratively with the IK functionality). Or, you will have to give the IK algorithm more intermediary points through which to pass.

Cheers

Schindler
Posts: 27
Joined: 16 Jul 2013, 09:59

Re: Oscillation of robot with masses

Post by Schindler »

thanks for your hints. I try to improve my scene this way.

Even if the robot has reached a certain point, the robot is not able to keep its position stable. This oscillations when the robot should stand still is are the biggest problem for me.

coppelia
Site Admin
Posts: 10339
Joined: 14 Dec 2012, 00:25

Re: Oscillation of robot with masses

Post by coppelia »

Try using a smaller damping value for the DLS resolution of your IK. And maybe also adjust the masses/inertias for better stability.

Cheers

Post Reply