Posts

paper in JMS & NAMRC

/in , /by

 

Anuj Gautam, Manish Raj Aryal, Sourabh Deshpande, Shailesh Padalkar, Mikhail Nikolaenko, Ming Tang, Sam Anand, IIoT-enabled digital twin for legacy and smart factory machines with LLM integration, Journal of Manufacturing Systems, Volume 80, 2025, Pages 511-523, ISSN 0278-6125

The paper is also published in the NAMRC 2025 conference.

Anuj Gautam , Manish Raj Aryal, Sourabh Deshpande, Shailesh Padalkar, Mikhail Nikolaenko, Ming Tang, Sam Anand. IIoT-enabled Digital Twin for legacy and smart factory machines with LLM integration. 53rd SME North American Manufacturing Research Conference (NAMRC), Clemson Univ. 06/2025.

 

Abstract

The recent advancement in Large Language Models (LLMs) has significantly transformed the field of natural data interpretation, translation, and user training. However, a notable gap exists when LLMs are tasked to assist with real-time context-sensitive machine data. The paper presents a multi-agent LLM framework capable of accessing and interpreting real-time and historical data through an Industrial Internet of Things (IIoT) platform for evidence-based inferences. The real-time data is acquired from several legacy machine artifacts (such as seven-segment displays, toggle switches, and knobs), smart machines (such as 3D printers), and building data (such as sound sensors and temperature measurement devices) through MTConnect data streaming protocol. Further, a multi-agent LLM framework that consists of four specialized agents – a supervisor agent, a machine-expertise agent, a data visualization agent, and a fault-diagnostic agent is developed for context-specific manufacturing tasks. This LLM framework is then integrated into a digital twin to visualize the unstructured data in real time. The paper also explores how LLM-based digital twins can serve as real time virtual experts through an avatar, minimizing reliance on traditional manuals or supervisor-based expertise. To demonstrate the functionality and effectiveness of this framework, we present a case study consisting of legacy machine artifacts and modern machines. The results highlight the practical application of LLM to assist and infer real-time machine data in a digital twin environment.

Human Digital Twin

/in /by

High-Fidelity Digital Human Modeling and Visualization

This research investigates advanced methods for constructing high-fidelity digital humans, with a focus on both technical innovation and applied use in immersive environments. The project integrates generative artificial intelligence, image-based modeling, and visualization pipelines to advance realism, interactivity, and usability in human–computer interaction.

Aim 1. Conversational Digital Agents Driven by Large Language Models (LLMs).
The first aim is to utilize large language models (LLMs) as the core engine for conversational digital humans. By embedding LLM-driven reasoning into virtual agents, the project seeks to create responsive, adaptive, and context-aware “talking agents.” These agents will simulate naturalistic dialogue, provide interactive guidance, and adapt to user needs across diverse scenarios such as education, healthcare training, and collaborative design.

Aim 2. Photorealistic Skin and Visual Fidelity Through Scanned Data.
The second aim focuses on the visual accuracy of digital humans. High-resolution image scans will be processed to reconstruct human skin with detailed fidelity, including surface textures, translucency, and micro-geometric variations. The resulting models are capable of 4K photorealistic rendering (click image to view sample output), significantly enhancing realism in simulation and visualization. This fidelity is crucial for applications where nuanced perception—such as empathy, trust, or attentiveness—depends on subtle visual cues.

Significance.
By combining intelligent conversational capabilities with photorealistic appearance, this research advances the next generation of digital humans. The outcomes will support applications in extended reality (XR), therapeutic and clinical training, collaborative design education, and digital twin environments, where authenticity of both interaction and appearance directly influences user engagement and effectiveness

High-fidelity digital human.

Read more

Digital Twin of Cincinnati

/in , /by

A realtime flythrough demo for Digital Twin of City Cincinnati

Digital Futures Building at the University of Cincinnati

Destroy Alien buildings near the UC campus. Project developed by students Cooper Pflaum and Nishanth Chidambaram. 

Digital Twin, LLM & IIOT

/in , , /by

IIOT for legacy and intelligent factory machines with XR and LLM feedback with a Digital Twin demonstration of real-time IOT for architecture/building applications using Omniverse. 

  • PIs: Sam Anand, Ming Tang.
  • Students: Anuj Gautama, Mikhail Nikolaenko, Ahmad Alrefai, Aayush Kumar, Manish Raj Aryal,c, Eian Bennett, Sourabh Deshpande 

$40,000. UC Industry 4.0/5.0 Institute Consortium Research Project: 01.2024-01.2025

The project centers on the development of a Digital Twin (DT) and a multi-agent Large Language Model (LLM) framework designed to access and interpret real-time and historical data through an Industrial Internet of Things (IIoT) platform. Real-time data is sourced from legacy machines and smart machines, integrating Building Information Modeling (BIM) with environmental sensors. The multi-agent LLM framework comprises specialized agents and supports diverse user interfaces, including screen-based systems, Virtual Reality (VR) environments, and mobile devices, enabling versatile interaction, data visualization, and analysis.

The research evaluates leading DT platforms—Autodesk Tandem, NVIDIA Omniverse, and Unreal Engine—focusing on their capabilities to integrate IoT and BIM data while supporting legacy machine systems.  Autodesk Tandem excelled in seamlessly combining BIM metadata with real-time IoT streams for building operations and system scalability.  NVIDIA Omniverse demonstrated unmatched rendering fidelity and collaborative features through its Universal Scene Description (USD) framework. Unreal Engine, notable for its immersive visualization, proved superior for LLM integration, leveraging 3D avatars and conversational AI to enhance user interaction.

Read more