mlsanityData to 7 Oct 2026

Constraint-Aware Diffusion Priors for High-Fidelity and Versatile Quadruped Locomotion

arXiv 2605.08804 ↗First posted 9 May 2026cs.ROcited by 1

On the map: Diffusion models → Diffusion and flow policies for robotics

also in 1 more
Abstract (arXiv)

Reinforcement learning combined with imitation learning has significantly advanced biomimetic quadrupedal locomotion. However, scaling these frameworks to massive, multi-source datasets exposes fundamental bottlenecks. First, traditional GAN-based discriminators are prone to mode collapse, struggling to capture diverse motion distributions from uncurated datasets. Second, existing kinematic priors suffer from out-of-distribution (OOD) tracking conflicts, leading to severe unintended heading drifts during complex maneuvers. Furthermore, deploying unconstrained priors to physical hardware poses critical safety risks by disregarding actuator dynamics. To overcome these challenges, we propose Diff-CAST (Diffusion-guided Constraint-Aware Symmetric Tracking), a novel motion prior framework leveraging the multi-modal distribution modeling capabilities of diffusion models for stylistic rewards. Diff-CAST effectively replaces traditional GAN discriminators, unlocking robust data scaling on heterogeneous collections. To ensure high-fidelity intent execution and reliable real-world deployment, we introduce a comprehensive Sim2Real architecture integrating Symmetric Augmented Command Conditioning (SACC) for drift-free tracking, and Constrained RL for hardware safety. Experiments on a quadruped demonstrate that Diff-CAST mitigates mode collapse, enables seamless transitions between diverse skills, and ensures robust, hardware-compliant locomotion.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Diffusion and flow policies for robotics, one of 14 directions in Diffusion models. Builds on 29 of its 34 references that are in the map. First posted on arXiv 9 May 2026.

Builds on

29 of 34 references.

Closest ancestors · 6

Same direction, newest first

  1. 2025Steering Your Diffusion Policy with Latent Space Reinforcement Learning18 Jun
  2. 2024Diffusion-Reward Adversarial Imitation Learning25 May
  3. 2023Diffusion Reward: Learning Rewards via Conditional Video Diffusion21 Dec
  4. 2023IDQL: Implicit Q-Learning as an Actor-Critic Method with Diffusion Policies20 Apr
  5. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusion7 Mar
  6. 2022SE(3)-DiffusionFields: Learning smooth cost functions for joint grasp and motion optimization through diffusion8 Sep

Other references · 23

  1. 2020Conservative Q-Learning for Offline Reinforcement Learningcited by 1,777
  2. 2022Human Motion Diffusion Modelcited by 1,062
  3. 2022Planning with Diffusion for Flexible Behavior Synthesiscited by 1,023
  4. 2020Out-of-Distribution Generalization via Risk Extrapolation (REx)cited by 737
  5. 2021RMA: Rapid Motor Adaptation for Legged Robotscited by 430
  6. 2019Learning agile and dynamic motor skills for legged robotscited by 402
  7. 2020Learning Agile Robotic Locomotion Skills by Imitating Animalscited by 229
  8. 2020Speech Gesture Generation from the Trimodal Context of Text, Audio, and Speaker Identitycited by 199
  9. 2022Walk These Ways: Tuning Robot Control for Generalization with Multiplicity of Behaviorcited by 113
  10. 2023ANYmal Parkour: Learning Agile Navigation for Quadrupedal Robotscited by 105
  11. 2023CALM: Conditional Adversarial Latent Models for Directable Virtual Characterscited by 91
  12. 2022Adversarial Motion Priors Make Good Substitutes for Complex Reward Functionscited by 68
11 more
  1. 2024DiffuseLoco: Real-Time Legged Locomotion Control with Diffusion from Offline Datasetscited by 47
  2. 2025Track Any Motions under Any Disturbancescited by 32
  3. 2022Learning Agile Skills via Adversarial Imitation of Rough Partial Demonstrationscited by 30
  4. 2022Safe Reinforcement Learning for Legged Locomotioncited by 29
  5. 2022Advanced Skills through Multiple Adversarial Motion Priors in Reinforcement Learningcited by 23
  6. 2024Symmetry Considerations for Learning Task Symmetric Robot Policiescited by 18
  7. 2024SoloParkour: Constrained Reinforcement Learning for Visual Locomotion from Privileged Experiencecited by 10
  8. 2025SPiDR: A Simple Approach for Zero-Shot Safety in Sim-to-Real Transfercited by 8
  9. 2025Learning Multi-Skill Legged Locomotion Using Conditional Adversarial Motion Priorscited by 2
  10. 2025BarlowWalk: Self-supervised Representation Learning for Legged Robot Terrain-adaptive Locomotioncited by 1
  11. 2025Coordinated Humanoid Robot Locomotion with Symmetry Equivariant Reinforcement Learning Policycited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Behavior Foundations for Quadruped Robots: ABot-C0 Technical Reportcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Offline Adaptation of Quadruped Locomotion using Diffusion Models
  2. 2024Preference Aligned Diffusion Planner for Quadrupedal Locomotion Control
  3. 2026Model-Based Geometry-Aware Generative Optimization for Constrained Locomotion Planning ↗
  4. 2026Unleashing Infinite Motion: Scaling Expressive Quadrupedal Motion via Generative Video Priors
  5. 2025Multi-Loco: Unifying Multi-Embodiment Legged Locomotion via Reinforcement Learning Augmented Diffusion
  6. 2025Integrating Diffusion-based Multi-task Learning with Online Reinforcement Learning for Robust Quadruped Robot Control
  7. 2026Learning Quadruped Walking from Seconds of Demonstration
  8. 2026Lagrangian Perturbation Diffusion Steering: Latent Reinforcement Learning for Generative Policies