A Grasp-based Motion Planning Algorithm for Character Animation

Abstract

The design of autonomous characters capable of planning their own motions continues to be a challenge for computer animation. We present a novel kinematic motion planning algorithm for character animation which addresses some of the outstanding problems. The problem domain for our algorithm is as follows: given a constrained environment with designated handholds and footholds, plan a motion through this space towards some desired goal. Our algorithm is based on a stochastic search procedure which is guided by a combination of geometric constraints, posture heuristics, and distance-togoal metrics. The method provides a single framework for the use of multiple modes of locomotion in planning motions through these constrained, unstructured environments. We illustrate our results with demonstrations of a human character using walking, swinging, climbing, and crawling in order to navigate through various obstacle courses.

Planning contact points for humanoid robots

“…The advantage of such a set of constraints over Eq. (8) is the possibility to take a subset of it. We use later the following weaker positioning constraints:…”

“…For such environments, several approaches proved to be efficient to plan motion for humanoid walking [1][2][3][4] or even running [5][6][7]. The seminal work in [8] illustrates motion generation from contact prints for general biped computer graphics figures. More recently impressive walking behaviors are generated automatically [9,10] even for highly uneven virtual terrains [11][12][13] (see also [14] in robotics).…”

State of the Art in Example‐Based Motion Synthesis for Virtual Characters in Interactive Applications

“…In this section, we give an overview of motion planning , a class of techniques for synthesis of optimal motion sequences that achieve complex high‐level goals. Motion planning techniques originate from the field of robotics and have been applied to a variety of motion tasks—locomotion (walking, running, climbing, crawling) of one or more characters [CLS03, KVDP01, KNK*01, LK05, LK06, PLS03, SKG05], grasping and manipulation of environment objects [YKH04, KKKL94], obstacle dodging [PLS03], etc. Unlike techniques that search for appropriate motion clips while using only local information, motion planning methods consider the entire relevant state space and generate motion sequences that are close to being globally optimal , that is they are near‐guaranteed to achieve objectives in the best, most expedient manner possible.…”

“…Motion planners which use PRMs are presented in [CLS03, SKG05, PLS03, KVDP01]. Choi et al [CLS03] construct a probabilistic roadmap by randomly sampling the state space for footprints.…”

Motion patches

“…With such a highdimensional configuration space, most optimal path planning algorithms are either computation-intensive for searching through configuration spaces exhaustively or memory-intensive for maintaining discretized configuration spaces. Many researchers used a lowdimensional configuration space (e.g., body and footprint locations) and randomized sampling techniques for producing character animation of locomotion, crawling, and climbing [Choi et al 2003;Kalisiak and van de Panne 2001;Kuffner et al 2001;Lau and Kuffner 2005;Pettre et al 2003;Sung et al 2005], and grasping and manipulating an object [Koga et al 1994;Yamane et al 2004]. Our motion patches can be thought of as a way to create discrete configuration spaces memory-efficient without losing the diversity and subtle details of captured motion data.…”