Latent Conditioned Loco-Manipulation Using Motion Priors

2025/09/19 by Stępień, Maciej, Rafael Kourdis, C. Roux +4 · 1 citation
Computer Science · Engineering · #FOS: Computer and information sciences #Human Motion and Animation #Human Pose and Action Recognition #Robotic Locomotion and Control #Robotics (cs.RO)

paper · pdf · doi:10.48550/arxiv.2509.16061

openalex publication_date 2025/09/19 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Although humanoid and quadruped robots provide a wide range of capabilities, current control methods, such as Deep Reinforcement Learning, focus mainly on single skills. This approach is inefficient for solving more complicated tasks where high-level goals, physical robot limitations and desired motion style might all need to be taken into account. A more effective approach is to first train a multipurpose motion policy that acquires low-level skills through imitation, while providing latent space control over skill execution. Then, this policy can be used to efficiently solve downstream tasks. This method has already been successful for controlling characters in computer graphics. In this work, we apply the approach to humanoid and quadrupedal loco-manipulation by imitating either simple synthetic motions or kinematically retargeted dog motions. We extend the original formulation to handle constraints, ensuring deployment safety, and use a diffusion discriminator for better imitation quality. We verify our methods by performing loco-manipulation in simulation for the H1 humanoid and Solo12 quadruped, as well as deploying policies on Solo12 hardware. Videos and code are available at https://gepetto.github.io/LaCoLoco/

Citations

• Humanoid Locomotion and Manipulation: Current Progress and Challenges in Control, Planning, and Learning
• HOVER: Versatile Neural Whole-Body Controller for Humanoid Robots
• Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes
• Diffusion-Reward Adversarial Imitation Learning
• Learning Visual Quadrupedal Loco-Manipulation from Demonstrations
• CaT: Constraints as Terminations for Legged Locomotion Reinforcement Learning
• Pedipulate: Enabling Manipulation Skills using a Quadruped Robot's Leg
• Reinforcement Learning for Versatile, Dynamic, and Robust Bipedal Locomotion Control
• Legs as Manipulator: Pushing Quadrupedal Agility Beyond Locomotion
• Versatile Skill Control via Self-supervised Adversarial Imitation of Unlabeled Mixed Motions
• Adversarial Motion Priors Make Good Substitutes for Complex Reward Functions
• Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement\n Learning
• Denoising Diffusion Probabilistic Models
• Mode-adaptive neural networks for quadruped motion control
• Diversity is All You Need: Learning Skills without a Reward Function
• Proximal Policy Optimization Algorithms
• Generative Adversarial Imitation Learning
• Generative Adversarial Networks
• On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming

Cited by

• Learning Reusable Hybrid Motion Priors for Humanoid Locomotion from Motion Imitation

Related

• Human Motion Control of Quadrupedal Robots using Deep Reinforcement Learning
• Learning Visual Quadrupedal Loco-Manipulation from Demonstrations
• Imitate and Repurpose: Learning Reusable Robot Movement Skills From Human and Animal Behaviors
• MLM: Learning Multi-task Loco-Manipulation Whole-Body Control for Quadruped Robot with Arm
• Deep Imitation Learning for Humanoid Loco-manipulation through Human Teleoperation