RESEARCH AREAS

AI-Native Digital Twins & Generative Simulation

We develop AI-native digital twins that can be generated and updated from natural language, CAD and layout data, sensor information, and execution traces. Our research combines LLM-based modeling, DEVS, hybrid simulation, and adaptive multi-fidelity techniques to build scalable simulation environments. 

• LLM-based simulation model generation

• DEVS and hybrid simulation

• Adaptive multi-fidelity simulation

Physical AI & Autonomous Manufacturing

We study Physical AI systems in which robots, AMRs, human workers, and AI agents interact within virtual manufacturing environments. Digital twins are used to learn, test, and validate autonomous decisions before their deployment to physical systems.

• Autonomous anomaly detection and response

• Multi-agent collaboration

• Isaac Sim and Sim-to-Real validation

Decision Intelligence, Optimization & Safety

We combine digital twins, AI, simulation, and optimization to support production scheduling, layout planning, resource allocation, AMR operations, and industrial safety. Real-time data and predictive models are used to identify risks and recommend effective operational decisions.

• Simulation-based optimization

• Deep reinforcement learning

• Scheduling, layout, and AMR optimization