Affiliation:
1. The University of British Columbia, Vancouver, BC, Canada
2. National University of Singapore
Abstract
The difficulty of developing control strategies has been a primary bottleneck in the adoption of physics-based simulations of human motion. We present a method for learning robust feedback strategies around given motion capture clips as well as the transition paths between clips. The output is a control graph that supports real-time physics-based simulation of multiple characters, each capable of a diverse range of robust movement skills, such as walking, running, sharp turns, cartwheels, spin-kicks, and flips. The control fragments that compose the control graph are developed using guided learning. This leverages the results of open-loop sampling-based reconstruction in order to produce state-action pairs that are then transformed into a linear feedback policy for each control fragment using linear regression. Our synthesis framework allows for the development of robust controllers with a minimal amount of prior knowledge.
Publisher
Association for Computing Machinery (ACM)
Subject
Computer Graphics and Computer-Aided Design
Cited by
82 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. Too Stiff, Too Strong, Too Smart;Proceedings of the ACM on Computer Graphics and Interactive Techniques;2023-08-16
2. MAAIP;Proceedings of the ACM on Computer Graphics and Interactive Techniques;2023-08-16
3. Composite Motion Learning with Task Control;ACM Transactions on Graphics;2023-07-26
4. Learning Physically Simulated Tennis Skills from Broadcast Videos;ACM Transactions on Graphics;2023-07-26
5. Bidirectional GaitNet: A Bidirectional Prediction Model of Human Gait and Anatomical Conditions;Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Proceedings;2023-07-23