My research aims to develop a novel AI backend for multi‑agent XR systems that can generate and adapt immersive industrial environments in real time from text, images, and videos, with the core objective of improving safety and decision‑making in safety‑critical, high‑risk workplaces.
It addresses key social and economic challenges such as workplace accidents, skills shortages, and operational inefficiencies in industrial domains by enabling intelligent XR-based guidance, adaptive training, and situational awareness that reduce human error and downtime. The research contributes to safer and more efficient industrial operations while supporting sustainable workforce development. Methodologically, it integrates LLMs for reasoning and task-adaptive instruction generation, computer vision for scene understanding and hologram registration, RAG for grounding domain knowledge, and deep reinforcement learning for real‑time environment and agent adaptation.
