Research
Founded in 2026 in the Department of Mechanical Engineering at the University of South Carolina, the IRIS Lab bridges manufacturing, cybersecurity, and data science to make industrial systems more secure, explainable, and resilient. South Carolina is home to the world’s largest BMW plant, Volvo and Mercedes-Benz assembly operations, Boeing’s 787 facility, and Scout Motors’ new electric-vehicle plant taking shape just outside Columbia — an industrial base whose security and quality challenges shape the lab’s agenda. The lab’s aim is to turn security-aware system design and explainable analytics into methods these manufacturers can put into practice.
The work is organized around three interrelated thrusts, each developing the scientific foundations and decision-support frameworks needed for secure, reliable, resilient, and sustainable smart manufacturing systems.
Secure-by-Design Manufacturing Systems
How do you catch an attack that slips past IT defenses and shows up only in what the machine produces?
Because cyber-physical attacks alter the physical process rather than just the network, they can compromise an automotive assembly line or a supplier’s machining cell without triggering conventional alarms. We build manufacturing-specific taxonomies and risk models that trace how such threats propagate across a production system, and we develop physics-informed defenses — from monitoring CNC power signals to protecting process data with privacy-preserving transformations — that catch and prevent what cyber-only tools miss.
Explainable Analytics for Manufacturing
Why did this defect occur — and can the model tell us?
In safety-critical production, a prediction is only useful if engineers can act on it. Rather than black-box models, we develop analytics grounded in domain ontologies and causal reasoning — including a conversational AI agent for defect diagnosis and Bayesian causal models of process–structure–property relationships — that explain why defects arise and support reliable decisions, from laser powder bed fusion to water treatment systems.
Resilient & Sustainable Process Design
Can we manufacture where there is no supply chain?
Whether on the lunar surface or on a factory floor without original design files, resilient manufacturing means working with what is available. We design resource-efficient materials and simulation-driven processes for constrained environments, including a hydrogel-based lunar regolith formulation for in-space 3D printing and a reverse-engineering framework that reconstructs additively manufactured parts from 3D scans to within 0.1% dimensional error when design files are unavailable.
Looking ahead, we're building toward a cyber-physical manufacturing security testbed, multimodal digital twins, and explainable machine-learning architectures for discrete and continuous manufacturing systems.