procedural foot placement

[Novakasa]

this might be the strategy:

• once a locked foot exceeds a certain distance (from hip position?) and the mapped velocity would increase that distance, unlock that foot
• continuously update the next target lock position while the foot is unlocked, probably have a fixed time of how long the foot is in the air
• the naive target position should be current position + foot offset + extrapolation based on current forces assuming the normal stays the same (extrapolate along plane). The actual position is then calculated with a raycast to this position.
• we probably need a mechanism so both feet are kind of operating on a cycle as well with opposite phases, not sure yet how this interacts with the first point