UrbanGround: From Local Perception to Spatial Agency in a Real-Scale City
Shanghai Jiao Tong University
Sandbox built from Hong Kong’s 3D geospatial data that supports first-person navigation for evaluating MLLM agents. Three research questions probe scene grounding, long-distance navigation, and robustness to environmental changes. MLLM agents handle basic visual recognition well but struggle with orientation, pedestrian awareness, and sustained goal-directed behavior over extended exploration.
Why it matters
70 upvotes on HF Daily Papers. Provides a real-scale, geo-faithful testbed for spatial MLLM agents rather than synthetic indoor scenes.
Importance: 2/5
official confirmation
Sources
official
arXiv listing