Page 1 of 1
Ragdoll Inverse Kinematics
Posted: Tue Feb 14, 2012 12:26 am
by Garold
Has anyone experimented with trying to make a ragdoll walk?
I envisage characters changing posture and stride depending on the terrain.
I have read this post
viewtopic.php?f=4&t=373&p=3993&hilit=in ... tics#p3993 but wondered if anyone has tried something.
Re: Ragdoll Inverse Kinematics
Posted: Tue Feb 14, 2012 12:57 am
by Norbo
Making a dynamic ragdoll walk and inverse kinematics are usually separated conceptually.
Moving a ragdoll involves motorizing the joints and perhaps adding a few more controls to compensate for the difficulty of actually making something balance dynamically. These motorized joints can then try to reach any goal whether it's defined by a traditional keyframed animation, inverse kinematics, or something else. I believe there have been a few people who have done something like this. [By the way, in the intervening years since that post, joint motors and all of that have been fully implemented.]
Inverse kinematics operates on goal states and transform chains. It can modify the character animation bones directly rather than working on a heavy physical ragdoll. You can get a lot fancier too by moving beyond traditional joint solving (e.g. planning future foot placement). Posture modification would probably be better suited to a special animation modification that is applied under certain circumstances as opposed to inverse kinematics, though.
Since the first concept actually deals with dynamics and physics, BEPUphysics is useful in implementing it. The second concept is in the realm of nonphysical animation, so there's not much that is built in that would help.