Puppets are highly dynamic hybrid systems, where each puppet executes a sequence of motions described in a high-level script, or play. The author of a puppet play describes the puppet performance by selecting different motions, such as “walk” or “dance,” and choreographing the motions with music and other puppets. This choreography introduces spatial and timing constraints in order to prevent puppets from colliding and tangling their strings. Our work modifies “standard” MDL so that it can properly describe puppet motion programs. We then use switched-time optimal control techniques to develop a MDL compiler that adjusts mode timing and “energy” scaling of these initial motion programs.
This work was funded by the NSF.