Research map / Multi-agent systems and games / Human-AI and goal-conditioned
MetaUrban: An Embodied AI Simulation Platform for Urban Micromobility
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
x
Problem
x
Approach
x
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Train on MetaUrban-train and evaluate on MetaUrban-unseen | PPO PointNav · PointNav unseen | SR | Helped |
| Fine-tune the MetaUrban-train pretrained model on MetaUrban-finetune instead of training directly on MetaUrban-finetune (ablation) | PPO · MetaUrban-unseen distribution | success rate | Helped |
| Increase generated training scenes from 5 to 1,000 (ablation) | PPO PointNav · PointNav | success rate: 12% to 46% | Helped |
| Increase static-object density (ablation) | PPO · PointNav train and test | SR | Hurt |
| Increase dynamic-agent density (ablation) | PPO · SocialNav train and test | SR | Hurt |
| Use Safe RL rather than standard RL | Navigation policies · PointNav and SocialNav | Cost | Helped |
| Reduce COCO max speed to 10 km/h (ablation) | COCO (base) · PointNav straight and intersection blocks | SR: 47% -> 62%; 41% -> 56% | Helped |
| Reduce COCO max steering angle to 10 degree (ablation) | COCO (base) · PointNav straight and intersection blocks | SR: 47% -> 17%; 41% -> 22% | Hurt |
| Use a conservative wheelchair design | Mobile-machine policy · PointNav intersection street block | Cost: 1.10 | Helped |
| Use an aggressive mobility-scooter design | Mobile-machine policy · PointNav intersection street block | Cost: 1.43 | Hurt |
Closest papers by meaning
- Towards Autonomous Micromobility through Scalable Urban Simulation (CVPR 2025 · similarity 0.93)
- Synthesizing Photorealistic and Dynamic Urban Environments for Multimodal Robot Navigation and Collaboration (NeurIPS 2025 · similarity 0.88)
- CityWalker: Learning Embodied Urban Navigation from Web-Scale Videos (CVPR 2025 · similarity 0.86)
- Learning to Drive in New Cities Without Human Demonstrations (NeurIPS 2026 · similarity 0.84)
- Habitat 3.0: A Co-Habitat for Humans, Avatars, and Robots (ICLR 2024 · similarity 0.84)
- CitySeeker: How Do VLMs Explore Embodied Urban Navigation with Implicit Human Needs? (ICLR 2026 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.