Immersive (VR/MR/XR) and Digital Twin-Driven Workforce Development
Immersive (VR/MR/XR) and Digital Twin-Driven Workforce Development
Our research question is how the expertise of experienced workers, both their work techniques and their safety judgment, can be extracted and transferred to new workers. Workforce development in this sense starts from evidence of how experts perform and decide and ends with a new worker who carries that performance into real work. We pursue the question in three connected areas. Wearable sensing and biomechanics capture how experienced workers move and load their bodies and the ergonomic risks they face. Applied AI turns that evidence and the governing safety rules into decision support. XR and digital twin environments then transfer that technique and judgment to new workers, increasingly in human-robot collaborative work.
Traditional instruction gives trainees little feedback on the postures and motions that separate an expert from a novice. We build mixed-reality training systems and digital twins that show new workers how experts perform a task and measure whether the skill holds on the physical task. Under an NSF EPSCoR Research Fellows award with the University of South Carolina as partner institution, our doctoral student built a mixed-reality training system and a digital twin of a collaborative robot in USC’s McNair Center and pilot-tested the training on a human-robot collaborative assembly task. In construction, we lead a West Virginia Division of Highways project on immersive workforce training and real-time wearable monitoring for road construction crews. Our review of extended reality in maintenance operations [J.16] and our task-sequencing framework for skill transfer [J.23] guide the training design.
Sponsors: NSF EPSCoR RII Track-4 Research Fellows (#2429709); NSF EPSCoR RII Track-2 FEC seed grant; WVDOT Division of Highways (RP-2025-18).
Selected papers: [J.23], [J.16], [J.07], [J.03], [C.22], [C.10].
AI-Enabled Risk Analysis and Decision Support
Safety judgment is the second part of expertise. We apply generative AI and natural language processing to the rules and records that govern industrial work. Advisee-led studies with a USC collaborator applied generative AI to OSHA compliance in manufacturing [J.24] and analyzed Industry 4.0 workforce competency requirements from job-market data [J.22]. A review of 87 studies on construction struck-by fatalities found that frontline roles defined in OSHA and ANSI/ASSP standards, such as spotters, receive little research attention [J.18]. In work-zone safety, we lead WorkZoneAI, a Division of Highways project to develop an AI and digital twin safety co-pilot for dynamic work-zone risk, and we are co-investigator on a U.S. Department of Transportation Safe Streets and Roads for All project on AI-assisted traffic safety analysis.
Sponsors: WVDOT Division of Highways (RP-2026-04, WorkZoneAI); U.S. DOT Safe Streets and Roads for All (Co-I).
Selected papers: [J.24], [J.22], [J.18], [J.13], [C.24].
Wearable-Based Ergonomic and Risk Analysis
Estimating physical load and fatigue in the field is the basis for both areas above. Our doctoral work developed a whole-body inverse dynamics method that estimates joint loads from inertial motion capture alone [J.06] and showed that expert masons use techniques that limit joint load while working more efficiently [J.03, J.08]. At WVU we extended these motion methods to multimodal physiological sensing: a single wrist-worn sensor combining heart rate, skin temperature, and movement jerk gives a validated fatigue assessment during manual material handling [J.19], and a wearable machine-learning study of healthcare workers’ physical fatigue received an IISE best paper award [C.21]. With doctoral advisees we study radiation protection and musculoskeletal risk in interventional radiology [J.21, J.20, J.14]. In a USDA AgrAbility project led by WVU Extension, the lab will run on-farm wearable risk assessments before and after assistive technology is installed for farmers with disabilities.
Sponsors: USDA NIFA AgrAbility (WVU Extension lead); WVU Research and Scholarship Advancement Grant.
Selected papers: [J.19], [J.21], [J.20], [J.14], [J.06], [J.04], [J.01], [C.21].
Foundations: Advanced Masonry Work Systems
The program began with assessment methods for advanced masonry work systems at the University of Waterloo. Motion capture and biomechanical analysis of 66 masons in four experience groups showed that experts with 20 or more years of experience limit the load on their joints while working more efficiently than less experienced masons [J.03]. Clustering their whole-body motion identified the expert postures for each phase of a block lift [J.07], and a force-assist pallet lowered peak lumbar compression at pick-up by about 40% while raising productivity by about 10% [J.04].
Selected papers: [J.08], [J.07], [J.05], [J.04], [J.03], [J.02], [C.08], [C.05].