AI RESEARCH
COBALT: Crowdsourcing Robot Learning via Cloud-Based Teleoperation with Smartphones
arXiv CS.AI
•
ArXi:2605.19138v1 Announce Type: cross The scarcity of large-scale, high-quality nstration data remains a bottleneck in scaling imitation learning for robotic manipulation. We present COBALT, a teleoperation platform designed to cratize robot learning at scale both in simulation and in the real world. By leveraging vectorized environments, our scalable, load-balanced infrastructure s concurrent teleoperation by multiple users on a single GPU, yielding a significant reduction in teleoperation cost.