Newest papers

DONIC: Generalist Dexterous Bimanual Tracking Control from Human Demonstrations

Mingze Sun, Ziang Cao, Yuan Fei, Qiankuan Tao, Zihan Wang, Tianqi Liu, Junhao Chen, Haiwen Diao, Fangzhou Hong, Zhaoxi Chen, Ziwei Liu · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29997 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

A pesar de los recientes avances en el seguimiento de referencia de objetos únicos, los modelos existentes se centran principalmente en la manipulación de manos únicas y un rango limitado de interacciones, dejando sin resolver los comportamientos de destreza bimanual comunes en la vida real. La escasez de datos de manipulación bimanual ricos en contacto y la creciente complejidad de los contactos de coordinación hacen que el desarrollo de un rastreador bimanual generalista sea particularmente desafiante. Por lo tanto, presentamos 1) DONIC, el primer rastreador bimanual generalista entrenado conjuntamente en categorías de operaciones y fuentes de movimiento humano, junto con 2) BiHOI, un conjunto de datos multi-visual de 1,575 secuencias que abarcan 229 objetos y enfatizan la interacción bimanual sostenida,

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

Abstract

Despite recent progress in dexterous reference tracking, existing models focus primarily on single-hand manipulation and a limited range of interactions, leaving the bimanual dexterous behaviors common in real life unaddressed. The scarcity of contact-rich bimanual manipulation data and the increased complexity of coordinating contacts make developing a generalist bimanual tracker particularly challenging. Therefore, we present 1) DONIC, the first generalist bimanual tracker jointly trained across operation categories and human-motion sources, together with 2) BiHOI, a multi-view dataset of 1,575 sequences spanning 229 objects and emphasizing sustained bimanual interaction, where reconstructed interactions are retargeted into robot references. DONIC combines proximal policy optimization with behavioral cloning in a feedback loop that recycles verified successful executions into training demonstrations. An object-centered per-hand reward maintains object-motion feedback while supervising each hand according to its reference contact role. Clip-Centered Advantages remove reference-dependent advantage offsets while preserving within-trajectory differences. Together, these designs enable one single policy to learn coordinated manipulation across heterogeneous references. Without finetuning, a single checkpoint achieves 86.79% success on unseen trajectories of training objects and 57.32% on held-out objects, surpassing the state-of-the-art bimanual tracking baselines by 77.36 and 45.73 percentage points.

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

Topics: Computer science, Artificial intelligence, Tracking (education), Generalist and specialist species, Robot, Control (management)

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