D-Robotics' Uranus simulates robot actions in a learned world
What happens when a robot takes an action? How will the objects move, and how will the environment respond?
Instead of testing every idea on physical hardware, you can explore what happens next in a learned world. Predicting how actions translate into motion, object interactions, and changes across multiple camera views.
That's action-controllable generation in Uranus, D-Robotics' data-driven robot simulation platform. It lets you try out candidate policies in a learned world first.
▶
✓ Verified
2026-10-09