A data-driven and theory-guided framework for developing and validating human-robot collaboration training modules for the construction workforce

Abstract

The growing integration of robotics and artificial intelligence in construction is reshaping project execution and creating new requirements for workforce capabilities in human-robot collaboration (HRC). However, construction organizations and academic institutions lack systematic and scalable approaches for translating HRC competencies into structured training systems aligned with workforce development and management needs. This study presents a data-driven and theory-guided framework for developing and validating HRC training modules to support construction workforce readiness. The framework integrates natural language processing-based competency augmentation, an ADDIE-driven instructional design process, and Delphi-based expert validation into a unified and reproducible workflow. An initial set of HRC competencies derived from prior literature was augmented using industry data, resulting in a validated framework of 50 HRC competencies across knowledge, skills, and abilities. These competencies were cognitively classified using Revised Bloom’s Taxonomy and translated into measurable learning objectives through a structured, theory-constrained pipeline. Seven training modules were developed to reflect progressive competency development, from foundational robotics knowledge to system-level reasoning, safety, and performance evaluation. A two-round Delphi study with academic and industry experts established consensus on the relevance of competencies, module-competency alignment, and cognitive classification, while identifying challenges in operationalizing higher-order and socio-cognitive competencies into measurable outcomes. The findings demonstrate the robustness and practical relevance of the proposed framework for workforce training design. This study contributes to engineering management by providing a scalable methodology for translating HRC competencies into validated training modules and by demonstrating that assessment requirements differ systematically across knowledge, skill, and ability competencies, thereby supporting workforce planning, training standardization, assessment design, and construction robotics adoption.