Abstract
A longstanding goal in character animation is to combine data-driven specification of behavior with a system that can execute a similar behavior in a physical simulation, thus enabling realistic responses to perturbations and environmental variation. We show that well-known reinforcement learning (RL) methods can be adapted to learn robust control policies capable of imitating a broad range of example motion clips, while also learning complex recoveries, adapting to changes in morphology, and accomplishing user-specified goals. Our method handles keyframed motions, highly-dynamic actions such as motion-captured flips and spins, and retargeted motions. By combining a motion-imitation objective with a task objective, we can train characters that react intelligently in interactive settings, e.g., by walking in a desired direction or throwing a ball at a user-specified target. This approach thus combines the convenience and motion quality of using motion clips to define the desired style and appearance, with the flexibility and generality afforded by RL methods and physics-based animation. We further explore a number of methods for integrating multiple clips into the learning process to develop multi-skilled agents capable of performing a rich repertoire of diverse skills. We demonstrate results using multiple characters (human, Atlas robot, bipedal dinosaur, dragon) and a large variety of skills, including locomotion, acrobatics, and martial arts.
Supplemental Material
- Shailen Agrawal, Shuo Shen, and Michiel van de Panne. 2013. Diverse Motion Variations for Physics-based Character Animation. Symposium on Computer Animation (2013). Google ScholarDigital Library
- Shailen Agrawal and Michiel van de Panne. 2016. Task-based Locomotion. ACM Trans. Graph. 35, 4, Article 82 (July 2016), 11 pages. Google ScholarDigital Library
- Marc G. Bellemare, Sriram Srinivasan, Georg Ostrovski, Tom Schaul, David Saxton, and Rémi Munos. 2016. Unifying Count-Based Exploration and Intrinsic Motivation. CoRR abs/1606.01868 (2016). arXiv:1606.01868Google Scholar
- Greg Brockman, Vicki Cheung, Ludwig Pettersson, Jonas Schneider, John Schulman, Jie Tang, and Wojciech Zaremba. 2016a. OpenAI Gym. CoRR abs/1606.01540 (2016). arXiv:1606.01540Google Scholar
- Greg Brockman, Vicki Cheung, Ludwig Pettersson, Jonas Schneider, John Schulman, Jie Tang, and Wojciech Zaremba. 2016b. OpenAI Gym. arXiv:arXiv:1606.01540Google Scholar
- Bullet. 2015. Bullet Physics Library, http://bulletphysics.org.Google Scholar
- Stelian Coros, Philippe Beaudoin, and Michiel van de Panne. 2009. Robust Task-based Control Policies for Physics-based Characters. ACM Trans. Graph. (Proc. SIGGRAPH Asia) 28, 5 (2009), Article 170. Google ScholarDigital Library
- Stelian Coros, Philippe Beaudoin, and Michiel van de Panne. 2010. Generalized Biped Walking Control. ACM Transctions on Graphics 29, 4 (2010), Article 130. Google ScholarDigital Library
- M. Da Silva, Y. Abe, and J. Popovic. 2008. Simulation of Human Motion Data using Short-Horizon Model-Predictive Control. Computer Graphics Forum (2008).Google Scholar
- Yan Duan, Xi Chen, Rein Houthooft, John Schulman, and Pieter Abbeel. 2016. Benchmarking Deep Reinforcement Learning for Continuous Control. CoRR abs/1604.06778 (2016). arXiv:1604.06778 Google ScholarDigital Library
- Justin Fu, John Co-Reyes, and Sergey Levine. 2017. EX2: Exploration with Exemplar Models for Deep Reinforcement Learning. In Advances in Neural Information Processing Systems 30. Curran Associates, Inc., 2574--2584.Google Scholar
- Sehoon Ha and C Karen Liu. 2014. Iterative training of dynamic skills inspired by human coaching techniques. ACM Transactions on Graphics 34, 1 (2014). Google ScholarDigital Library
- Perttu Hämäläinen, Joose Rajamäki, and C Karen Liu. 2015. Online control of simulated humanoids using particle belief propagation. ACM Transactions on Graphics (TOG) 34, 4 (2015), 81. Google ScholarDigital Library
- Nicolas Heess, Dhruva TB, Srinivasan Sriram, Jay Lemmon, Josh Merel, Greg Wayne, Yuval Tassa, Tom Erez, Ziyu Wang, S. M. Ali Eslami, Martin A. Riedmiller, and David Silver. 2017. Emergence of Locomotion Behaviours in Rich Environments. CoRR abs/1707.02286 (2017). arXiv:1707.02286Google Scholar
- Nicolas Heess, Gregory Wayne, Yuval Tassa, Timothy P. Lillicrap, Martin A. Riedmiller, and David Silver. 2016. Learning and Transfer of Modulated Locomotor Controllers. CoRR abs/1610.05182 (2016). arXiv:1610.05182Google Scholar
- Jonathan Ho and Stefano Ermon. 2016. Generative Adversarial Imitation Learning. In Advances in Neural Information Processing Systems 29. Curran Associates, Inc., 4565--4573. Google ScholarDigital Library
- Daniel Holden, Taku Komura, and Jun Saito. 2017. Phase-functioned Neural Networks for Character Control. ACM Trans. Graph. 36, 4, Article 42 (July 2017), 13 pages. Google ScholarDigital Library
- Daniel Holden, Jun Saito, and Taku Komura. 2016. A Deep Learning Framework for Character Motion Synthesis and Editing. ACM Trans. Graph. 35, 4, Article 138 (July 2016), 11 pages. Google ScholarDigital Library
- Rein Houthooft, Xi Chen, Yan Duan, John Schulman, Filip De Turck, and Pieter Abbeel. 2016. Curiosity-driven Exploration in Deep Reinforcement Learning via Bayesian Neural Networks. CoRR abs/1605.09674 (2016). arXiv:1605.09674Google ScholarDigital Library
- Yoonsang Lee, Sungeun Kim, and Jehee Lee. 2010a. Data-driven Biped Control. In ACM SIGGRAPH 2010 Papers (SIGGRAPH '10). ACM, New York, NY, USA, Article 129, 8 pages. Google ScholarDigital Library
- Yoonsang Lee, Moon Seok Park, Taesoo Kwon, and Jehee Lee. 2014. Locomotion Control for Many-muscle Humanoids. ACM Trans. Graph. 33, 6, Article 218 (Nov. 2014), 11 pages. Google ScholarDigital Library
- Yongjoon Lee, Kevin Wampler, Gilbert Bernstein, Jovan Popović, and Zoran Popović. 2010b. Motion Fields for Interactive Character Locomotion. In ACM SIGGRAPH Asia 2010 Papers (SIGGRAPH ASIA '10). ACM, New York, NY, USA, Article 138, 8 pages. Google ScholarDigital Library
- Sergey Levine, Jack M. Wang, Alexis Haraux, Zoran Popović, and Vladlen Koltun. 2012. Continuous Character Control with Low-Dimensional Embeddings. ACM Transactions on Graphics 31, 4 (2012), 28. Google ScholarDigital Library
- Libin Liu and Jessica Hodgins. 2017. Learning to Schedule Control Fragments for Physics-Based Characters Using Deep Q-Learning. ACM Trans. Graph. 36, 3, Article 29 (June 2017), 14 pages. Google ScholarDigital Library
- Libin Liu, Michiel van de Panne, and KangKang Yin. 2016. Guided Learning of Control Graphs for Physics-Based Characters. ACM Transactions on Graphics 35, 3 (2016). Google ScholarDigital Library
- Libin Liu, KangKang Yin, Michiel van de Panne, Tianjia Shao, and Weiwei Xu. 2010. Sampling-based Contact-rich Motion Control. ACM Transctions on Graphics 29, 4 (2010), Article 128. Google ScholarDigital Library
- Josh Merel, Yuval Tassa, Dhruva TB, Sriram Srinivasan, Jay Lemmon, Ziyu Wang, Greg Wayne, and Nicolas Heess. 2017. Learning human behaviors from motion capture by adversarial imitation. CoRR abs/1707.02201 (2017). arXiv:1707.02201Google Scholar
- Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A. Rusu, Joel Veness, Marc G. Bellemare, Alex Graves, Martin Riedmiller, Andreas K. Fidjeland, Georg Ostrovski, Stig Petersen, Charles Beattie, Amir Sadik, Ioannis Antonoglou, Helen King, Dharshan Kumaran, Daan Wierstra, Shane Legg, and Demis Hassabis. 2015. Human-level control through deep reinforcement learning. Nature 518, 7540 (Feb. 2015), 529--533.Google ScholarCross Ref
- Igor Mordatch, Emanuel Todorov, and Zoran Popović. 2012. Discovery of Complex Behaviors Through Contact-invariant Optimization. ACM Trans. Graph. 31, 4, Article 43 (July 2012), 8 pages. Google ScholarDigital Library
- Uldarico Muico, Yongjoon Lee, Jovan Popović, and Zoran Popović. 2009. Contact-aware nonlinear control of dynamic characters. In ACM Transactions on Graphics (TOG), Vol. 28. ACM, 81. Google ScholarDigital Library
- Ashvin Nair, Bob McGrew, Marcin Andrychowicz, Wojciech Zaremba, and Pieter Abbeel. 2017. Overcoming Exploration in Reinforcement Learning with Demonstrations. CoRR abs/1709.10089 (2017). arXiv:1709.10089Google Scholar
- Xue Bin Peng, Glen Berseth, and Michiel van de Panne. 2015. Dynamic Terrain Traversal Skills Using Reinforcement Learning. ACM Trans. Graph. 34, 4, Article 80 (July 2015), 11 pages. Google ScholarDigital Library
- Xue Bin Peng, Glen Berseth, and Michiel van de Panne. 2016. Terrain-Adaptive Locomotion Skills Using Deep Reinforcement Learning. ACM Transactions on Graphics (Proc. SIGGRAPH 2016) 35, 4 (2016). Google ScholarDigital Library
- Xue Bin Peng, Glen Berseth, KangKang Yin, and Michiel van de Panne. 2017a. DeepLoco: Dynamic Locomotion Skills Using Hierarchical Deep Reinforcement Learning. ACM Transactions on Graphics (Proc. SIGGRAPH 2017) 36, 4 (2017). Google ScholarDigital Library
- Xue Bin Peng, Michiel van de Panne, and KangKang Yin. 2017b. Learning Locomotion Skills Using DeepRL: Does the Choice of Action Space Matter?. In Proc. ACM SIGGRAPH / Eurographics Symposium on Computer Animation. Google ScholarDigital Library
- Aravind Rajeswaran, Sarvjeet Ghotra, Sergey Levine, and Balaraman Ravindran. 2016. EPOpt: Learning Robust Neural Network Policies Using Model Ensembles. CoRR abs/1610.01283 (2016). arXiv:1610.01283Google Scholar
- Aravind Rajeswaran, Vikash Kumar, Abhishek Gupta, John Schulman, Emanuel Todorov, and Sergey Levine. 2017. Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations. CoRR abs/1709.10087 (2017). arXiv:1709.10087Google Scholar
- Alla Safonova and Jessica K Hodgins. 2007. Construction and optimal search of interpolated motion graphs. In ACM Transactions on Graphics (TOG), Vol. 26. ACM, 106. Google ScholarDigital Library
- John Schulman, Sergey Levine, Philipp Moritz, Michael I. Jordan, and Pieter Abbeel. 2015a. Trust Region Policy Optimization. CoRR abs/1502.05477 (2015). arXiv:1502.05477 Google ScholarDigital Library
- John Schulman, Philipp Moritz, Sergey Levine, Michael I. Jordan, and Pieter Abbeel. 2015b. High-Dimensional Continuous Control Using Generalized Advantage Estimation. CoRR abs/1506.02438 (2015). arXiv:1506.02438Google Scholar
- John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. 2017. Proximal Policy Optimization Algorithms. CoRR abs/1707.06347 (2017). arXiv:1707.06347Google Scholar
- Dana Sharon and Michiel van de Panne. 2005. Synthesis of Controllers for Stylized Planar Bipedal Walking. In Proc. of IEEE International Conference on Robotics and Animation.Google Scholar
- Kwang Won Sok, Manmyung Kim, and Jehee Lee. 2007. Simulating biped behaviors from human motion data. In ACM Transactions on Graphics (TOG), Vol. 26. ACM, 107. Google ScholarDigital Library
- R. Sutton, D. Mcallester, S. Singh, and Y. Mansour. 2001. Policy Gradient Methods for Reinforcement Learning with Function Approximation. , 1057--1063 pages. Google ScholarDigital Library
- Richard S. Sutton and Andrew G. Barto. 1998. Introduction to Reinforcement Learning (1st ed.). MIT Press, Cambridge, MA, USA. Google ScholarDigital Library
- Jie Tan, Karen Liu, and Greg Turk. 2011. Stable Proportional-Derivative Controllers. IEEE Comput. Graph. Appl. 31, 4 (2011), 34--44. Google ScholarDigital Library
- Yuval Tassa, Tom Erez, and Emanuel Todorov. 2012. Synthesis and stabilization of complex behaviors through online trajectory optimization. In Intelligent Robots and Systems (IROS), 2012 IEEE/RSJ International Conference on. IEEE, 4906--4913.Google ScholarCross Ref
- Yee Whye Teh, Victor Bapst, Wojciech Marian Czarnecki, John Quan, James Kirkpatrick, Raia Hadsell, Nicolas Heess, and Razvan Pascanu. 2017. Distral: Robust Multitask Reinforcement Learning. CoRR abs/1707.04175 (2017). arXiv:1707.04175Google Scholar
- Kevin Wampler, Zoran Popović, and Jovan Popović. 2014. Generalizing Locomotion Style to New Animals with Inverse Optimal Regression. ACM Trans. Graph. 33, 4, Article 49 (July 2014), 11 pages. Google ScholarDigital Library
- Jack M. Wang, Samuel R. Hamner, Scott L. Delp, Vladlen Koltun, and More Specifically. 2012. Optimizing locomotion controllers using biologically-based actuators and objectives. ACM Trans. Graph (2012). Google ScholarDigital Library
- Ronald J. Williams. 1992. Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine Learning 8, 3 (01 May 1992), 229--256. Google ScholarDigital Library
- Jungdam Won, Jongho Park, Kwanyu Kim, and Jehee Lee. 2017. How to Train Your Dragon: Example-guided Control of Flapping Flight. ACM Trans. Graph. 36, 6, Article 198 (Nov. 2017), 13 pages. Google ScholarDigital Library
- Yuting Ye and C Karen Liu. 2010a. Optimal feedback control for character animation using an abstract model. In ACM Transactions on Graphics (TOG), Vol. 29. ACM, 74. Google ScholarDigital Library
- Yuting Ye and C. Karen Liu. 2010b. Synthesis of Responsive Motion Using a Dynamic Model. Computer Graphics Forum 29, 2 (2010), 555--562.Google ScholarCross Ref
- KangKang Yin, Kevin Loken, and Michiel van de Panne. 2007. SIMBICON: Simple Biped Locomotion Control. ACM Trans. Graph. 26, 3 (2007), Article 105. Google ScholarDigital Library
Index Terms
- DeepMimic: example-guided deep reinforcement learning of physics-based character skills
Recommendations
SFV: reinforcement learning of physical skills from videos
Data-driven character animation based on motion capture can produce highly naturalistic behaviors and, when combined with physics simulation, can provide for natural procedural responses to physical perturbations, environmental changes, and ...
C·ASE: Learning Conditional Adversarial Skill Embeddings for Physics-based Characters
SA '23: SIGGRAPH Asia 2023 Conference PapersWe present C · ASE, an efficient and effective framework that learns Conditional Adversarial Skill Embeddings for physics-based characters. C · ASE enables the physically simulated character to learn a diverse repertoire of skills while providing ...
Learning Physically Simulated Tennis Skills from Broadcast Videos
We present a system that learns diverse, physically simulated tennis skills from large-scale demonstrations of tennis play harvested from broadcast videos. Our approach is built upon hierarchical models, combining a low-level imitation policy and a high-...
Comments