Features
Multi-Modal Annotations
Detailed visual, kinematic, environmental, and behavioral labels per robot instance — enabling advanced detection, classification, and motion prediction models.
Robot-Centric Metadata
Each robot is linked to structured metadata covering platform type, country of origin, sensor loadout, payload configuration, and connectivity traits — ideal for simulation, intent modeling, or coordination systems.
Real-World Context Coverage
Curated from diverse public video sources across indoor, urban, industrial, and open-field scenarios — supporting robust perception models that generalize to real deployment conditions.
