mlsanityData to 14 Oct 2026

ASAP: Aligning Simulation and Real-World Physics for Learning Agile Humanoid Whole-Body Skills

arXiv 2502.01143 ↗First posted 3 Feb 2025cs.RO cs.AI cs.LGcited by 130

Not on the map. Why?

Cited by 40 in Deep reinforcement learning methods, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Humanoid robots hold the potential for unparalleled versatility in performing human-like, whole-body skills. However, achieving agile and coordinated whole-body motions remains a significant challenge due to the dynamics mismatch between simulation and the real world. Existing approaches, such as system identification (SysID) and domain randomization (DR) methods, often rely on labor-intensive parameter tuning or result in overly conservative policies that sacrifice agility. In this paper, we present ASAP (Aligning Simulation and Real-World Physics), a two-stage framework designed to tackle the dynamics mismatch and enable agile humanoid whole-body skills. In the first stage, we pre-train motion tracking policies in simulation using retargeted human motion data. In the second stage, we deploy the policies in the real world and collect real-world data to train a delta (residual) action model that compensates for the dynamics mismatch. Then, ASAP fine-tunes pre-trained policies with the delta action model integrated into the simulator to align effectively with real-world dynamics. We evaluate ASAP across three transfer scenarios: IsaacGym to IsaacSim, IsaacGym to Genesis, and IsaacGym to the real-world Unitree G1 humanoid robot. Our approach significantly improves agility and whole-body coordination across various dynamic motions, reducing tracking error compared to SysID, DR, and delta dynamics learning baselines. ASAP enables highly agile motions that were previously difficult to achieve, demonstrating the potential of delta action learning in bridging simulation and real-world dynamics. These results suggest a promising sim-to-real direction for developing more expressive and agile humanoids.

Key facts. Cited by 130 papers in the map (data through 14 Oct 2026). Builds on 86 of its 107 references that are in the map. First posted on arXiv 3 Feb 2025.

Builds on

86 of 107 references.

References · 86

  1. 2017Proximal Policy Optimization Algorithmscited by 15,215
  2. 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,498
  3. 2021Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learningcited by 998
  4. 2017Sim-to-Real Transfer of Robotic Control with Dynamics Randomizationcited by 672
  5. 2020Learning Quadrupedal Locomotion over Challenging Terraincited by 604
  6. 2021Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learningcited by 512
  7. 2021RMA: Rapid Motor Adaptation for Legged Robotscited by 508
  8. 2019Learning agile and dynamic motor skills for legged robotscited by 504
  9. 2022Learning robust perceptive locomotion for quadrupedal robots in the wildcited by 382
  10. 2018Sim-to-Real: Learning Agile Locomotion For Quadruped Robotscited by 319
  11. 2023Orbit: A Unified Simulation Framework for Interactive Robot Learning Environmentscited by 295
  12. 2018Residual Reinforcement Learning for Robot Controlcited by 274
38 more
  1. 2020Learning Agile Robotic Locomotion Skills by Imitating Animalscited by 251
  2. 2023Perpetual Humanoid Control for Real-time Simulated Avatarscited by 218
  3. 2018Closing the Sim-to-Real Loop: Adapting Simulation Randomization with Real World Experiencecited by 217
  4. 2024OmniH2O: Universal and Dexterous Human-to-Humanoid Whole-Body Teleoperation and Learningcited by 189
  5. 2023Real-World Humanoid Locomotion with Reinforcement Learningcited by 187
  6. 2024HumanPlus: Humanoid Shadowing and Imitation from Humanscited by 185
  7. 2022Rapid Locomotion via Reinforcement Learningcited by 168
  8. 2018Residual Policy Learningcited by 164
  9. 2017Preparing for the Unknown: Learning a Universal Policy with Online System Identificationcited by 141
  10. 2024Expressive Whole-Body Control for Humanoid Robotscited by 139
  11. 2024Learning Human-to-Humanoid Real-Time Whole-Body Teleoperationcited by 130
  12. 2022Deep Whole-Body Control: Learning a Unified Policy for Manipulation and Locomotioncited by 121
  13. 2022In-Hand Object Rotation via Rapid Motor Adaptationcited by 118
  14. 2021Reinforcement Learning for Robust Parameterized Locomotion Control of Bipedal Robotscited by 116
  15. 2024MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpaintingcited by 116
  16. 2024TRAM: Global Trajectory and Motion of 3D Humans from in-the-wild Videoscited by 106
  17. 2023Synthesizing Physical Character-Scene Interactionscited by 104
  18. 2024Reinforcement Learning for Versatile, Dynamic, and Robust Bipedal Locomotion Controlcited by 99
  19. 2024Humanoid Parkour Learningcited by 96
  20. 2024HOVER: Versatile Neural Whole-Body Controller for Humanoid Robotscited by 91
  21. 2024From Imitation to Refinement -- Residual RL for Precise Assemblycited by 90
  22. 2022Concurrent Training of a Control Policy and a State Estimator for Dynamic and Robust Legged Locomotioncited by 88
  23. 2024ExBody2: Advanced Expressive Humanoid Whole-Body Controlcited by 88
  24. 2018Neural Lander: Stable Drone Landing Control using Learned Dynamicscited by 85
  25. 2019BayesSim: adaptive domain randomization via probabilistic inference for robotics simulatorscited by 85
  26. 2022Neural-Fly Enables Rapid Learning for Agile Flight in Strong Windscited by 85
  27. 2023PhysHOI: Physics-Based Imitation of Dynamic Human-Object Interactioncited by 76
  28. 2023Teach a Robot to FISH: Versatile Imitation from One Minute of Demonstrationscited by 75
  29. 2024Humanoid Locomotion as Next Token Predictioncited by 69
  30. 2023DreamWaQ: Learning Robust Quadrupedal Locomotion With Implicit Terrain Imagination via Deep Reinforcement Learningcited by 66
  31. 2024WoCoCo: Learning Whole-Body Humanoid Control with Sequential Contactscited by 61
  32. 2024TRANSIC: Sim-to-Real Policy Transfer by Learning from Online Correctioncited by 59
  33. 2019Sim-to-Real Transfer for Biped Locomotioncited by 56
  34. 2024Advancing Humanoid Locomotion: Mastering Challenging Terrains with Denoising World Model Learningcited by 56
  35. 2022Adapting Rapid Motor Adaptation for Bipedal Robotscited by 54
  36. 2024Mobile-TeleVision: Predictive Motion Priors for Humanoid Whole-Body Controlcited by 51
  37. 2024Learning Humanoid Locomotion with Perceptive Internal Modelcited by 50
  38. 2021Auto-Tuned Sim-to-Real Transfercited by 49

50 most cited of 86 All 86 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 59; 2026: 71.

130 papers.

Sort:

By area:

  1. 2025BeyondMimic: From Motion Tracking to Versatile Humanoid Control via Guided Diffusioncited by 165
  2. 2025SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Controlcited by 101
  3. 2025OmniRetarget: Interaction-Preserving Data Generation for Humanoid Whole-Body Loco-Manipulation and Scene Interactioncited by 71
  4. 2025KungfuBot: Physics-Based Humanoid Whole-Body Control for Learning Highly-Dynamic Skillscited by 52
  5. 2025TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenizationcited by 51
  6. 2025ResMimic: From General Motion Tracking to Humanoid Whole-body Loco-Manipulation via Residual Learningcited by 41
  7. 2025WholeBodyVLA: Towards Unified Latent VLA for Whole-Body Loco-Manipulation Controlcited by 36
  8. 2025SPIDER: Scalable Physics-Informed Dexterous Retargetingcited by 32
  9. 2025HuB: Learning Extreme Humanoid Balancecited by 28
  10. 2025Humanoid Everyday: A Comprehensive Robotic Dataset for Open-World Humanoid Manipulationcited by 25
  11. 2025TD-JEPA: Latent-predictive Representations for Zero-Shot Reinforcement Learningcited by 22
  12. 2025PhysHSI: Towards a Real-World Generalizable and Natural Humanoid-Scene Interaction Systemcited by 18
38 more
  1. 2025From Language to Locomotion: Retargeting-free Humanoid Control via Motion Latent Guidancecited by 18
  2. 2025From Experts to a Generalist: Toward General Whole-Body Control for Humanoid Robotscited by 17
  3. 2025CBF-RL: Safety Filtering Reinforcement Learning in Training with Control Barrier Functionscited by 17
  4. 2025Learning Sim-to-Real Humanoid Locomotion in 15 Minutescited by 16
  5. 2025DreamControl: Human-Inspired Whole-Body Humanoid Control for Scene Interaction via Guided Diffusioncited by 15
  6. 2025Opening the Sim-to-Real Door for Humanoid Pixel-to-Action Policy Transfercited by 15
  7. 2025Reconstructing 4D Spatial Intelligence: A Surveycited by 14
  8. 2026HumanX: Toward Agile and Generalizable Humanoid Interaction Skills from Human Videoscited by 14
  9. 2025KungfuBot2: Learning Versatile Motion Skills for Humanoid Whole-Body Controlcited by 13
  10. 2025MOSPA: Human Motion Generation Driven by Spatial Audiocited by 12
  11. 2025RL from Physical Feedback: Aligning Large Motion Models with Humanoid Controlcited by 11
  12. 2025DemoHLM: From One Demonstration to Generalizable Humanoid Loco-Manipulationcited by 11
  13. 2026ZEST: Zero-shot Embodied Skill Transfer for Athletic Robot Controlcited by 11
  14. 2025Neural Brain: A Neuroscience-inspired Framework for Embodied Agentscited by 10
  15. 2025MoRE: Mixture of Residual Experts for Humanoid Lifelike Gaits Learning on Complex Terrainscited by 10
  16. 2025OXE-AugE: A Large-Scale Robot Augmentation of OXE for Scaling Cross-Embodiment Policy Learningcited by 10
  17. 2026FRoM-W1: Towards General Humanoid Whole-Body Control with Language Instructionscited by 10
  18. 2025SkillBlender: Towards Versatile Humanoid Whole-Body Loco-Manipulation via Skill Blendingcited by 9
  19. 2025DexNDM: Closing the Reality Gap for Dexterous In-Hand Rotation via Joint-Wise Neural Dynamics Modelcited by 9
  20. 2025Humanoid Goalkeeper: Learning from Position Conditioned Task-Motion Constraintscited by 9
  21. 2026ULTRA: Unified Multimodal Control for Autonomous Humanoid Whole-Body Loco-Manipulationcited by 9
  22. 2025Towards Adaptable Humanoid Control via Adaptive Motion Trackingcited by 8
  23. 2025Olaf: Bringing an Animated Character to Life in the Physical Worldcited by 8
  24. 2026ZeroWBC: Learning Natural Whole-Body Humanoid Interaction from Human Egocentric Datacited by 8
  25. 2026FAWAM: Force-Aware World Action Models for Closed-Loop Contact-Rich Manipulationcited by 8
  26. 2025Natural Humanoid Robot Locomotion with Generative Motion Priorcited by 7
  27. 2025SLAC: Safe and Efficient Real-Robot Reinforcement Learning via Unsupervised Simulation Pre-Trainingcited by 7
  28. 2025Robot Trains Robot: Automatic Real-World Policy Adaptation and Learning for Humanoidscited by 7
  29. 2025GentleHumanoid: Learning Upper-body Compliance for Contact-rich Human and Object Interactioncited by 7
  30. 2025RoboMirror: Understand Before You Imitate for Video to Humanoid Locomotioncited by 7
  31. 2025MOVE: Motion-Guided Few-Shot Video Object Segmentationcited by 6
  32. 2025UrbanNav: Learning Language-Guided Urban Navigation from Web-Scale Human Trajectoriescited by 6
  33. 2026Learning Whole-Body Human-Humanoid Interaction from Human-Human Demonstrationscited by 6
  34. 2026ω-0: A Latent Predictive World Action Model for Concurrent Humanoid Loco-Manipulationcited by 6
  35. 2025Efficient Imitation under Misspecificationcited by 5
  36. 2025Motion Priors Reimagined: Adapting Flat-Terrain Skills for Complex Quadruped Mobilitycited by 5
  37. 2025InterAgent: Physics-based Multi-agent Command Execution via Diffusion on Interaction Graphscited by 5
  38. 2025CVP: Central-Peripheral Vision-Inspired Multimodal Model for Spatial Reasoningcited by 5

50 most cited of 130 All 130 (JSON)