Digital Twin of Skills (DToS): A Semantic Framework for Deterministic Human–Robot Orchestration in Industry 5.0

Abstract

Hybrid human-robot collaboration in Industry 5.0 calls for coordination that treats human and robotic resources as comparable contributors to shared objectives. Yet skills, capabilities, and execution conditions are still represented through fragmented, executor-specific models, with no common representation under which heterogeneous executors can be evaluated and constrained, and their selection justified. This article proposes the Digital Twin of Skills (DToS), a conceptual and standards-oriented framework that formalises a skill as an agent-agnostic entity defined by its intent, interface, operational context, and configurable constraints, and instantiated per executor (R-DToS for robotic, H-DToS for human resources). Rather than a static description, the DToS maintains a time-resolved executability state of each skill, updated from the executor’s condition as it changes; for human executors a parallel Human Safety Envelope (HSE) turns dynamic state variables into admissibility constraints. Requirements are derived from open industrial standards (IEC 62264, Asset Administration Shell), and the framework is realised as an executable reference implementation instantiated for two manufacturing pilots, furniture painting and textile classification for circular economy, where representative scenarios provide an initial, internally consistent feasibility check. The contribution is a semantic layer producing auditable, reconstructible verdicts, a deterministic basis for orchestrating operations and resources that also supports operator and equipment monitoring and the anticipation of admissibility windows. The framework reframes the central question from who can perform a task to who can do so safely, appropriately, and with traceable justification at a given moment.