In the videos below you can see some preliminary tests on applying sinusoidally sequenced torque to the parts that compose the Jacobian Model. There is a “bug” in the way that the torque is applied. Instead of applying it to the body parts in relation to the servo rotation axis, the torques are applied in relation to an axis that is at the center of mass of the body parts. Obviously this is not really the way things work with a multi-servo robotic construct in the real world. The torque in the real world is applied by the servo along the rotation axis.
The good news is that convex triangle mesh based collision seems to work.
Jacobian robot simulation with a sinusoidal signal applied to the legs:
A froggy Jacobian?
Actuatedcharacter's Blog 
July 15, 2010 at 2:59 pm |
your robot will move faster backwards than forwards. i can prove it.
July 15, 2010 at 6:39 pm |
We haven’t quite started with thinking about direction of movement but can you please elaborate on how you can prove that these legs can move faster in one direction than another?