TERRA: Terrain-Aware Reconstruction, Retargeting and Control for Musculoskeletal Locomotion

Hugging Face Daily Papers Papers
musculoskeletal-modeling reinforcement-learning locomotion motion-retargeting terrain-awareness robotics

Summary

TERRA is an end-to-end pipeline for terrain-aware reconstruction, retargeting, and control of musculoskeletal locomotion, enabling a single muscle-actuated RL policy trained on 9.4 hours of motion data to traverse diverse non-flat terrain with fewer anatomical and interaction violations.

Recent advances in musculoskeletal modeling and reinforcement learning have enabled muscle-actuated agents to reproduce increasingly complex human motions. Yet these capabilities remain largely confined to flat ground, in part because motion datasets rarely include aligned terrain geometry and because retargeting terrain interactions to complex musculoskeletal bodies is challenging. We present TERRA, an end-to-end pipeline for terrain-aware retargeting and control of musculoskeletal locomotion. From kinematic trajectories alone, TERRA combines terrain priors, estimated contacts, and negative free-space evidence to recover task-relevant support geometry. TERRA further considers anatomical, tendon-continuity, and contact constraints during retargeting. Using the resulting motion-terrain pairs from five datasets, we successfully train a single muscle-actuated control policy on 9.4 hours of diverse locomotion. Across reconstruction, retargeting, and held-out tracking benchmarks, TERRA improves terrain accuracy, sharply reduces anatomical and interaction violations, and achieves the highest observed completion rate over supported terrain families. Overall, TERRA provides a practical route from scene-less motion data to muscle-actuated locomotion over diverse non-flat terrain. Project website: https://cnai.epfl.ch/terra/
Original Article
View Cached Full Text

Cached at: 10/01/26, 04:23 PM

Paper page - TERRA: Terrain-Aware Reconstruction, Retargeting and Control for Musculoskeletal Locomotion

Source: https://huggingface.co/papers/2609.38653

Abstract

Recentadvancesinmusculoskeletalmodelingandreinforcementlearninghaveenabledmuscle-actuatedagentstoreproduceincreasinglycomplexhumanmotions.Yetthesecapabilitiesremainlargelyconfinedtoflatground,inpartbecausemotiondatasetsrarelyincludealignedterraingeometryandbecauseretargetingterraininteractionstocomplexmusculoskeletalbodiesischallenging.WepresentTERRA,anend-to-endpipelineforterrain-awareretargetingandcontrolofmusculoskeletallocomotion.Fromkinematictrajectoriesalone,TERRAcombinesterrainpriors,estimatedcontacts,andnegativefree-spaceevidencetorecovertask-relevantsupportgeometry.TERRAfurtherconsidersanatomical,tendon-continuity,andcontactconstraintsduringretargeting.Usingtheresultingmotion-terrainpairsfromfivedatasets,wesuccessfullytrainasinglemuscle-actuatedcontrolpolicyon9.4hoursofdiverselocomotion.Acrossreconstruction,retargeting,andheld-outtrackingbenchmarks,TERRAimprovesterrainaccuracy,sharplyreducesanatomicalandinteractionviolations,andachievesthehighestobservedcompletionrateoversupportedterrainfamilies.Overall,TERRAprovidesapracticalroutefromscene-lessmotiondatatomuscle-actuatedlocomotionoverdiversenon-flatterrain.Projectwebsite:https://cnai.epfl.ch/terra/

View arXiv pageView PDFProject pageGitHub6Add to collection

Get this paper in your agent:

hf papers read 2609\.38653

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.38653 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.38653 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.38653 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Geometric Action Model for Robot Policy Learning

Hugging Face Daily Papers

The Geometric Action Model (GAM) repurposes a pretrained geometric foundation model (GFM) as a unified backbone for language-conditioned robot manipulation, achieving higher accuracy, robustness, and efficiency than existing foundation-model-scale baselines across simulation and real-world benchmarks.

Transferability for General Reasoning: An Automated Curriculum for Multi-Domain RLVR

Hugging Face Daily Papers

This paper proposes Transfer-Aware Curriculum (TAC), a bandit-style online curriculum for multi-domain RLVR that prioritizes domains whose updates benefit other domains using gradient-geometry alignment. TAC improves macro-averaged accuracy on Qwen3-1.7B and Llama3.2-3B over fixed and learnability-only curricula.