I have always wanted a keyboard controlled, fully physically simulated powered-ragdoll that would walk, run, jump, stand using forces and torques (just like a real world person). With the accessibility of deep learning APIs, this appears to be becoming possible.
The current version can stand, walk, run, getup (not very robust yet), jump (do we really need to be able to land the jump?). Still very much a work in progress prototype, but it is almost usable. This version uses separate brains trained to stand/walk, run, jump, getup-prone, getup-supine. The figure switches brains depending on the situation.
Here is the same CC from a few weeks earlier:
Here are some of the learning to jump attempts. Unfortunately it is trying to stay in the air as long as possible, hence it is avoiding landing on its feet. Working on fixing that:
