How do I find my servo and battery size?

[Sarah Naeem]

Hi all

We are trying to understand the working of the Stanford pupper but are unable to figure out the dynamic modeling of it. We are currently employing Langrange approach. The calculations are getting too complicated so if anyone has a simpler solution to go about this or can help us with the Langrange approach, please do reach out. 

Regards

[Nathan]

Hi Sarah, if you're looking to generate the equations of motion by hand then no matter what method you use they will be quite complicated. Although I personally haven't derived symbolic dynamics expressions for the 12dof robot (and I haven't seen them written out in any papers), my guess is that the simplest method would be to use generalized coordinates to model each leg separately as if it was floating in space, and then use Newton/Euler equations to 1) model the body in 6DOF and 2) derive equations for the constraint forces holding the four legs to the body. The lagrange method is not well suited to modeling the whole robot together because the quaternion orientation of the floating base (robot body) makes the kinetic and potential energy calculations really really complex. Kane's method is better if going this route. If numerical dynamics are ok, then CRBA or ABA are the most common algorithms. Check out https://scaron.info/robotics/forward-dynamics.html for more info. Hope this helps!