Planning biped locomotion using motion capture data and probabilistic roadmaps

Abstract

Typical high-level directives for locomotion of human-like characters are useful for interactive games and simulations as well as for off-line production animation. In this thesis, we present a new scheme for planning natural-looking locomotion of a biped figure with three high-level directives to facilitate rapid motion prototyping and task-level motion generation for interactive applications: The first directive is to generate the motion of a character from a given start configuration to a goal configuration with a set of live-captured motion clips, the second one is to guide the character through an ordered set of configurations in sequence, and the last one is to make the character follow a user-specified trajectory within a tolerance. Based on a novel combination of probabilistic path planning and hierarchical displacement mapping, our scheme consists of three parts: roadmap construction, roadmap search, and motion generation. We randomly sample a set of valid footholds of the biped figure from the environment to construct a directed graph, called a roadmap, which guides the locomotion of the figure. Every edge of the roadmap is associated with a live-captured motion clip. Augmenting the roadmap with a posture transition graph, we traverse it to obtain the sequence of input motion clips and that of target footprints in accordance with a given high-level directive. The target footprints are further refined to yield a better resulting motion. They are also post-processed under interactive user guides. Finally, we adapt the motion sequence to the constraints specified by the footprint sequence to generate a desired locomotion. Provided with a rich set of motion clips, our scheme enables a human-like character to move over uneven terrain with a variety of motions such as running on flat terrain, walking over stepping stones, and jumping over a moat, walking up/down a stairway and hurdling an obstacle.