Exploring the Effectiveness of GAN-based Approach and Reinforcement Learning in Character Boxing Task
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
For decades, creating a desired locomotive motion in a goal-oriented manner has been a challenge in character animation. Datadriven methods using generative models have demonstrated efficient ways of predicting long sequences of motions without the need for explicit conditioning. While these methods produce high-quality long-term motions, they can be limited when it comes to synthesizing motion for challenging novel scenarios, such as punching a random target. A state-of-the-art solution to overcome this limitation is by using a GAN Discriminator to imitate motion data clips and incorporating reinforcement learning to compose goaloriented motions. In this paper, our research aims to create characters performing combat sports such as boxing, using a novel reward design in conjunction with existing GAN-based approaches. We experimentally demonstrate that both the Adversarial Motion Prior [3] and Adversarial Skill Embeddings [4] methods are capable of generating viable motions for a character punching a random target, even in the absence of mocap data that specifically captures the transition between punching and locomotion. Also, with a single learned policy, multiple task controllers can be constructed through the TimeChamber framework.
Abstract
For decades, creating a desired locomotive motion in a goal-oriented manner has been a challenge in character animation. Datadriven methods using generative models have demonstrated efficient ways of predicting long sequences of motions without the need for explicit conditioning. While these methods produce high-quality long-term motions, they can be limited when it comes to synthesizing motion for challenging novel scenarios, such as punching a random target. A state-of-the-art solution to overcome this limitation is by using a GAN Discriminator to imitate motion data clips and incorporating reinforcement learning to compose goaloriented motions. In this paper, our research aims to create characters performing combat sports such as boxing, using a novel reward design in conjunction with existing GAN-based approaches. We experimentally demonstrate that both the Adversarial Motion Prior [3] and Adversarial Skill Embeddings [4] methods are capable of generating viable motions for a character punching a random target, even in the absence of mocap data that specifically captures the transition between punching and locomotion. Also, with a single learned policy, multiple task controllers can be constructed through the TimeChamber framework.
참고문헌 (31)
- L. Kovar, "Motion graphs", ACM SIGGRAPH 2008 classes, pp.1~10 (2008)
- J. Ho, "Generative adversarial imitation learning", Advances in neural information processing systems, 29 (2016)
- X. B. Peng, "Amp: Adversarial motion priors for stylized physics-based character control", ACM Transactions on Graphics (TOG), 40(4), pp.1~20 (2021)
- X. B. Peng, "ASE: Large-Scale Reusable Adversarial Skill Embeddings for Physically Simulated Characters", arXiv preprint arXiv:2205.01906, ~ (2022)
- D. Holden, "Phase-functioned neural networks for character control", ACM Transactions on Graphics (TOG), 36(4), pp.1~13 (2017)
- H. Zhang, "Mode-adaptive neural networks for quadruped motion control", ACM Transactions on Graphics(TOG), 37(4), pp.1~11 (2018)
- D. Holden, "Learned motion matching", ACM Transactions on Graphics (TOG), 39(4), pp.53:1~53:12 (2020)
- X. B. Peng, "Deeploco : Dynamic locomotion skills using hierarchical deep reinforcement learning", ACM Transactions on Graphics(TOG), 36(4), pp.1~13 (2017)
- W. Yu, "Learning symmetric and lowenergy locomotion", ACM Transactions on Graphics (TOG), 37(4), pp.1~12 (2018)
- A. Elgammal, "Human Motion", pp.25~56 (2008), Springer
- D. Holden, "A deep learning framework for character motion synthesis and editing", ACM Transactions on Graphics (TOG), 35(4), pp.1~11 (2016)
- H. Y. Ling, "Character controllers using motion vaes", ACM Transactions on Graphics (TOG), 39(4), pp.40:1~40:12 (2020)
- J. Won, "Physics-based character controllers using conditional VAEs", ACM Transactions on Graphics(TOG), 41(4), pp.1~12 (2022)
- H. Yao, "ControlVAE : Model-Based Learning of Generative Controllers for Physics-Based Characters", ACM Transactions on Graphics(TOG), 41(6), pp.1~16 (2022)
- G. E. Henter, "Moglow : Probabilistic and controllable motion synthesis using normalising flows", ACM Transactions on Graphics(TOG), 39(6), pp.1~14 (2020)
- J. Juravsky, "PADL: Language-Directed Physics-Based Character Control", 1~9
- S. Agrawal, "Task-based locomotion", ACM Transactions on Graphics(TOG), 35(4), pp.1~11 (2016)
- K. Lee, "Interactive character animation by learning multi-objective control", ACM Transactions on Graphics(TOG), 37(6), pp.1~10 (2018)
- J. Merel, "Catch & carry : reusable neural controllers for vision-guided whole-body tasks", ACM Transactions on Graphics (TOG), 39(4), pp.39:1~39:12 (2020)
- L. Fussell, "Supertrack : Motion tracking for physically simulated characters using supervised learning", ACM Transactions on Graphics(TOG), 40(6), pp.1~13 (2021)
- T. Bansal, "Emergent complexity via multi-agent competition", arXiv preprint arXiv : 1710. 03748, ~ (2017)
- B. Baker, "Emergent tool use from multi-agent autocurricula", arXiv preprint arXiv : 1909. 07528, ~ (2019)
- J. Won, "Control strategies for physically simulated characters performing two-player competitive sports", ACM Transactions on Graphics(TOG), 40(4), pp.1~11 (2021)
- "TimeChamber:A Massively Parallel Large Scale Self-Play Framework"[URL]
- CMU, "CMU Graphics Lab Motion Capture Database", ~ (2002)
- E. Coumans, "Bullet physics library", Open source:bulletphysics. org, 15(49), pp.5~ (2013)
- V. Makoviychuk, "Isaac gym: High performance gpubased physics simulation for robot learning", arXiv preprint arXiv:2108.10470, ~ (2021)
- G. Tevet, "Human motion diffusion model", arXiv preprint arXiv:2209.14916, ~ (2022)
- M. Zhang, "Motiondiffuse: Text-driven human motion generation with diffusion model", arXiv preprint arXiv:2208.15001, ~ (2022)
- Y. Shafir, "Human Motion Diffusion as a Generative Prior", arXiv preprint arXiv:2303.01418, ~ (2023)
- Y. Yuan, "PhysDiff: Physics-Guided Human Motion Diffusion Model", arXiv preprint arXiv:2212.02500, ~ (2022)