A Three-Layer Taxonomy and Predictive Digital Twin for Role-Aware Human–Robot Collaboration in 3D Concrete Printing

Abstract: 3D concrete printing (3DCP) has reached early industrial deployment, yet current standards exclude environmental, health, and safety aspects of robotic operation, and existing collaborative-robot frameworks do not address process-coupled hazards or multi-role construction environments. This paper proposes a three-layer taxonomy (Human, Robot, Site) and a site-wide predictive digital twin with five-second update cadence and short-horizon forecasting. The framework is operationalized through seven role-aware interaction patterns aligned with ISO 10218. A scenario-based evaluation of the Human-layer physiological pipeline using a proxy dataset demonstrates the feasibility of stress classification and calibration. The work provides a reference architecture for human–robot–site integration in 3DCP.