State-annotated motion graphs

Motion graphs have gained popularity in recent years as a means for re-using motion capture data by connecting previously unrelated segments of a recorded library. Current techniques for controlling movement of a character via motion graphs have largely focused on path planning which is difficult due to the density of connections found on the graph. We introduce "state-annotated motion graphs," a novel technique which allows high-level control of character behavior by using a dual representation consisting of both a motion graph and a behavior state machine. This special motion graph is generated from labeled data and then bound to a finite state machine with similar labels. At run-time, character behavior is simply controlled by switching states. We show that it is possible to generate rich, controllable motion without the need for deep planning. We demonstrate that, when applied to an interactive fighting testbed, simple state-switching controllers may be coded intuitively to create various effects.