# ADP: Adversarial Dynamics Priors for Physically Grounded Humanoid Locomotion

arXiv: https://arxiv.org/abs/2607.03454 · HTML: https://mlsanity.com/paper/2607.03454/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 32 references that are in the map.
- First posted on arXiv 3 Jul 2026.

## Builds on (most cited first, 22 of 22)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Learning Quadrupedal Locomotion over Challenging Terrain](https://mlsanity.com/paper/2010.11251/) (2020)
- [Learning robust perceptive locomotion for quadrupedal robots in the wild](https://mlsanity.com/paper/2201.08117/) (2022)
- [BeyondMimic: From Motion Tracking to Versatile Humanoid Control via Guided Diffusion](https://mlsanity.com/paper/2508.08241/) (2025)
- [ASAP: Aligning Simulation and Real-World Physics for Learning Agile Humanoid Whole-Body Skills](https://mlsanity.com/paper/2502.01143/) (2025)
- [Adversarial Motion Priors Make Good Substitutes for Complex Reward Functions](https://mlsanity.com/paper/2203.15103/) (2022)
- [Visual Imitation Enables Contextual Humanoid Control](https://mlsanity.com/paper/2505.03729/) (2025)
- [Highly Dynamic Quadruped Locomotion via Whole-Body Impulse Control and Model Predictive Control](https://mlsanity.com/paper/1909.06586/) (2019)
- [Generalizable Humanoid Manipulation with 3D Diffusion Policies](https://mlsanity.com/paper/2410.10803/) (2024)
- [Opt2Skill: Imitating Dynamically-feasible Whole-Body Trajectories for Versatile Humanoid Loco-Manipulation](https://mlsanity.com/paper/2409.20514/) (2024)
- [HumanMimic: Learning Natural Locomotion and Transitions for Humanoid Robot via Wasserstein Adversarial Imitation](https://mlsanity.com/paper/2309.14225/) (2023)
- [Model-free Reinforcement Learning for Robust Locomotion using Demonstrations from Trajectory Optimization](https://mlsanity.com/paper/2107.06629/) (2021)
- [DemoHLM: From One Demonstration to Generalizable Humanoid Loco-Manipulation](https://mlsanity.com/paper/2510.11258/) (2025)
- [ZEST: Zero-shot Embodied Skill Transfer for Athletic Robot Control](https://mlsanity.com/paper/2602.00401/) (2026)
- [Integrating Model-Based Footstep Planning with Model-Free Reinforcement Learning for Dynamic Legged Locomotion](https://mlsanity.com/paper/2408.02662/) (2024)
- [Trajectory Optimization Through Contacts and Automatic Gait Discovery for Quadrupeds](https://mlsanity.com/paper/1607.04537/) (2016)
- [Fast Trajectory Optimization for Legged Robots using Vertex-based ZMP Constraints](https://mlsanity.com/paper/1705.10313/) (2017)
- [Computationally-Robust and Efficient Prioritized Whole-Body Controller with Contact Constraints](https://mlsanity.com/paper/1807.01222/) (2018)
- [Fast Whole-Body Motion Control of Humanoid Robots with Inertia Constraints](https://mlsanity.com/paper/2007.06275/) (2020)
- [Robust Whole-Body Motion Control of Legged Robots](https://mlsanity.com/paper/1703.02326/) (2017)
- [Dynamic Complementarity Conditions and Whole-Body Trajectory Optimization for Humanoid Robot Locomotion](https://mlsanity.com/paper/2207.03198/) (2022)
- [DynaFlow: Dynamics-embedded Flow Matching for Physically Consistent Motion Generation from State-only Demonstrations](https://mlsanity.com/paper/2509.19804/) (2025)

## Built on by (most cited first, 1 of 1)

- [OmniMimic: Dynamics-completed Motion Augmentation for Multi-style Omnidirectional Quadruped Locomotion](https://mlsanity.com/paper/2609.20566/) (2026)

## Abstract (arXiv)

In this paper, we propose Adversarial Dynamics Priors (ADP) for perturbation-resilient humanoid locomotion control. Existing motion prior-based methods induce natural motion styles by imitating kinematic motion features, but they do not directly regularize dynamics features, such as CoM motion, centroidal momentum, contact forces, and contact states. To address this limitation, we replace kinematic motion-style feature with selected dynamics features extracted from locomotion trajectories as the target of adversarial regularization. To this end, we use trajectory optimization to construct a reference dataset and train a discriminator to evaluate whether policy-induced temporal windows are consistent with the resulting reference distribution. Without explicit motion tracking, ADP encourages policy rollouts to remain close to the reference support, even after perturbations. Experimental results show that, compared with AMP, the strongest baseline in our evaluation, ADP improves the 80%-success impulse threshold (J80) by 16.7%, while reducing direction-averaged recovery time and velocity tracking error by 47.9% and 35.4%, respectively.