EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

Towards Spatially Aware AI Agents for Pre-Existing 3D Engines Using Multimodal Reasoning

Dragoș Silion, Ciprian Amariței, Adrian Iftene

Computer Science Journal of Moldova · 2026-11 · 34 (2) · 183-201

DOI: 10.56415/csjm.v34.08 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computer Science Journal of Moldova, 2026-11. Verfasst von Dragoș Silion und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We introduce a modular framework powered by the Model Context Protocol (MCP) that enables agentic personas to operate inside 3D environments by perceiving the surrounding areas through both a physics engine collision system and visual information. We propose a two-layer vision system that can handle both 360 equirectangular inputs and focused mono-directional images. By pairing the system with a custom toolset of computer vision techniques optimized for the two types of input imagery, we enable agents to seamlessly perceive, navigate, and operate inside complex 3D environments.

Zitieren

APA

Silion, D., Amariței, C., & Iftene, A. (2026). Towards Spatially Aware AI Agents for Pre-Existing 3D Engines Using Multimodal Reasoning. Computer Science Journal of Moldova, 34(2), 183-201. https://doi.org/10.56415/csjm.v34.08

Vancouver

Silion D, Amariței C, Iftene A. Towards Spatially Aware AI Agents for Pre-Existing 3D Engines Using Multimodal Reasoning. Computer Science Journal of Moldova. 2026;34(2):183-201. doi:10.56415/csjm.v34.08

BibTeX
@article{Silion2026,
  author = {Silion, Dragoș and Amariței, Ciprian and Iftene, Adrian},
  title = {Towards Spatially Aware AI Agents for Pre-Existing 3D Engines Using Multimodal Reasoning},
  journal = {Computer Science Journal of Moldova},
  year = {2026},
  volume = {34},
  number = {2},
  pages = {183--201},
  doi = {10.56415/csjm.v34.08}
}
Kostenlos starten →