automated human motion in constrained environments maciej kalisiak [email protected]

31
Automated human motion in constrained environments Maciej Kalisiak [email protected]

Upload: hugo-dalton

Post on 13-Jan-2016

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Automated human motion in constrained environments

Maciej [email protected]

Page 2: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Introduction

• human character animation• constrained environments• kinematic method• currently 2D, extendible • sample solution

Page 3: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Path Planning

• piano mover’s problem

• given: start and goal configurations

• find connecting path

Page 4: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Application to Human Motion

Page 5: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Approach

• starting point: RPP

• additions:– moving while in contact with environment– notion of comfort– knowledge of human gaits

Page 6: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Understanding RPP

• Randomized Path Planning

• a path planning algorithm

Page 7: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Simplest “Planner”

• character’s state: q

• repeated perturbations,i.e., Brownian motion

• repeat until goal reached

Page 8: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

• discretize into grid

• potential = Manhattan distance to goal

• flood-fill

Building a Potential Field

Page 9: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Gradient Descent

• character point mass• sample q’s neighbourhood• pick sample with largest

drop in potential• iterate until goal reached• not feasible analytically

Page 10: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Local Minima

• gradient descent stops at any minimum

• use random walks to escape– Brownian motion of predetermined duration

• use backtracking if minimum too deep– revert to a previous point in solution,

followed by a random walk

Page 11: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Deep Minimum Example

Page 12: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Smoothing

• solution embodies complete history of search process

• also very noisy

• a trajectory filter post-process is applied– removes extraneous motion segments– makes remaining motion more fluid

Page 13: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Modifications

• grasps and grasp invariants

• comfort heuristic system

• gait finite state machine

• grasp-aware gradient descent, random walk, smoothing filters

Page 14: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Character Structure

• 10 links

• 9 joints

• 12 DOFs

• frequent re-rooting

Page 15: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Grasp Points

• represent potential points of contact

• three types

• reduce the grasp search space

• summarize surface characteristics

Page 16: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Grasp Invariants

• each gait dictates:– the number of grasps– the types of grasps

• enforced by the GFSM

• rest of planner must not alter existing grasps

Page 17: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Motion without Heuristics

Page 18: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Heuristic System

• each heuristic measures some quality of q

• D(q): overall discomfort, a potential field

• getting comfy: gradient descent through D(q)

Page 19: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Implemented Heuristics

Page 20: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

The Gait FSM

• states represent gaits

• each edge has:– geometric preconditions– motion recipe– priority

• self-loops: gait-preserving motion that changes grasps

Page 21: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Complete System

Page 22: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

More Results

Page 23: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Future Work

• 3D

• quadrupeds, other characters

• “grasp surfaces”

• non-limb grasping

• add concept of time, speed

• use machine learning

Page 24: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

~FIN~

http://www.dgp.toronto.edu/~mac/thesis

Page 25: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Appendix

(extra slides)

Page 26: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Alternate gradient descent view

Page 27: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Smoothing Algorithm

Page 28: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Need for Limb Smoothing

Page 29: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Limb Smoothing Solution

Page 30: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Implemented GFSM

Page 31: Automated human motion in constrained environments Maciej Kalisiak mac@dgp.toronto.edu

Contributions

• human character animation algorithmfor constrained environments– grasp point discretization of environment– grasp constraint– comfort modeling using heuristics– gait FSM– adapted RPP algorithms to grasp constraint