Search papers, labs, and topics across Lattice.
This paper introduces OASIS, a simulation-data-driven framework that enhances humanoid loco-manipulation by reconstructing realistic object assets from real-world images using a 3D generative model. By collecting trajectories through teleoperation in simulation and augmenting them with diverse domain randomizations, OASIS enables the training of a hierarchical visuomotor policy. The results demonstrate that this policy achieves superior success rates in real-world tasks compared to those trained solely on real-robot teleoperation data, highlighting the advantages of simulation in capturing varied environmental conditions.
OASIS shows that simulation can outperform real-world teleoperation in humanoid manipulation tasks by leveraging diverse environmental variations.
Recent progress in robot manipulation has been largely driven by learning from large-scale demonstrations. For humanoid robot loco-manipulation tasks, however, existing data sources force an unsatisfying tradeoff between trajectory quality and scalability. Real-world teleoperation provides the highest-quality trajectories but requires dedicated physical space and time-consuming scene resets. Simulation offers an alternative way out of this dilemma: it can produce clean, embodiment-aligned data at scale without any physical hardware. In this paper, we propose OASIS, a simulation-data-driven framework for humanoid loco-manipulation. OASIS automatically reconstructs realistic object assets from real-world images using a 3D generative model. Based on these assets, trajectories are first collected through teleoperation in simulation, and then augmented under diverse domain randomizations in a post-processing stage. With the resulting simulation data, we further design a hierarchical visuomotor policy for humanoid loco-manipulation. Extensive experiments on the real humanoid robot show that, under zero-shot deployment, the policy trained on our simulation data achieves higher success rates on most tasks than that trained on real-robot teleoperation data, owing largely to the broad lighting and environmental variations covered by our simulation rendering, which real-robot data fails to capture. The project page is available at https://oasis-humanoid.github.io/.
