Newest papers

Fast Motion Retargeting Pipeline for Scalable Humanoid Robot Learning

Daniel Marew, Sangjun Park, Nisal Perera, Donghyun Kim · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28470 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

Enquanto os conjuntos de dados de captura de movimento humano em larga escala fornecem uma rica fonte de demonstrações para o aprendizado de robôs humanoides, o retargeting dos vários movimentos para diferentes robôs humanoides continua a ser um desafio não trivial. Os tubos de otimização de trajetória cinodinâmica existentes são computacionalmente caros e muitas vezes exigem ajustes manuais por movimento. Nós apresentamos o FRDV (Retargeting Rápido com Validação Dinâmica), um quadro de transferência escalável consistindo em retargeting cinemático restrito rápido e um pipeline de validação de dinâmica adicional. Seu estágio cinemático integra um solucionador de programação quadrática sequencial personalizado, iteração em tempo real e processamento baseado em pedaços paralelos,

Machine translation (Meta M2M100); the English original follows.

Abstract

While large-scale human motion capture datasets provide a rich source of demonstrations for humanoid robot learning, retargeting the diverse motions to different humanoid robots remains a nontrivial challenge. Existing kinodynamic trajectory optimization pipelines are computationally expensive and often require manual adjustments per motion. We present FRDV (Fast Retargeting with Dynamics Validation), a scalable framework consisting of rapid constrained kinematic retargeting and an additional dynamics validation pipeline. Its kinematic stage integrates a custom Sequential Quadratic Programming solver, Real-Time Iteration, and parallel chunk-based processing, achieving up to 9,156 frames per second, approximately up to 136 times faster than GMR and ProtoMotions, while robot-specific constraints are incorporated into the optimization to reduce common retargeting artifacts, such as ground penetration and self-collision. For motions requiring additional verification, a contact implicit iLQR controller is utilized to test the retargeted reference under full rigid body dynamics. This dynamics-based validation identifies dynamically infeasible motions and refines contact-related artifacts such as foot floating and foot skating. We evaluate retargeting quality by training motion tracking reinforcement learning policies using several state-of-the-art pipelines and comparing their tracking fidelity and success rates. Furthermore, sim-to-real transfer is demonstrated on a Unitree G1 humanoid. Videos and code are available at: https://daroslab.github.io/frdv/.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Retargeting, Computer science, Artificial intelligence, Computer vision, Humanoid robot, Kinematics

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange