Enhancing complex technical environment training with immersive digital twins

Immersive digital twins hold great promise for advancing training and education in complex technical domains, offering interactive, spatially rich alternatives to traditional instruction. In semiconductor cleanrooms—environments where access is restricted and procedures are intricate—novices often rely on passive materials such as manuals and orientation videos, which do little to foster spatial understanding or procedural fluency. While recent photorealistic reconstruction techniques, such as NeRFs, have improved visual fidelity, they struggle to scale to large environments and do not readily support interactive simulation, thereby limiting their effectiveness for experiential learning. This thesis investigates whether an immersive digital twin with continuous photorealistic rendering and integrated environmental simulation can effectively support users in acquiring spatial and functional knowledge of a complex technical environment. To facilitate the investigation, we developed a realistic digital twin pipeline that constructs a cleanroom digital twin using 3D Gaussian Splatting (3DGS) for enhanced visual realism. This pipeline was integrated into Unity, incorporating a structural skeleton and dynamic simulations of airflow, temperature, pressure, and humidity. We conducted a mixed-methods evaluation consisting of a pre-test/post-test user study with 24 students unfamiliar with cleanroom environments, as well as expert interviews with nine graduate researchers and one cleanroom engineer. Results show significant learning gains among users and positive feedback from experts regarding the system’s realism and training value. These findings provide empirical evidence that immersive digital twins, when designed with both visual and functional fidelity, can enhance learning outcomes in complex technical environments.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois Urbana-Champaign

Year dc:date

2025

Author dc:creator

• Wei, Jionghao

Contributors dc:contributor

• Caesar, Matthew

• Immersive Digital Twin
• 3d Gaussian Splatting
• Computer Vision
• Scene Reconstruction
• Iot
• Cleanroom Simulation

Statement dc:rights

• All Rights Reserved.

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/130232

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F130232