[From Bill Powers (920710.0900)]
Avery Andrews (920709.2112) --
In the report she demonstrates a qualitative method for flipping a
plastic egg in a frying pan using an MIT/Utah hand on a Puma. There is
no configuration space planning, no stable grasp determination, and no
CAD model of the either the pan or egg. Instead there are a bunch of
phases which are either fixed action patterns or compliant moves.
Transitions between phases are signalled by changes in tendon tensions.
What does it do if there isn't actually any egg in the frying pan?