Objective
Robot dog autonomously navigates a simulated chemical spill, identifies casualties via face recognition, assesses scenes with VLM, and streams video to a remote operator.
Three Demo Scenes
- Scene 1: Office — unconscious workers
- Scene 2: Stacked Boxes — trapped worker
- Scene 3: Chemical Spill — source identification
Key Capabilities
LLM-powered planning, Real-time video streaming, Face recognition, VLM scene analysis, Two-way speech