Towards a solution to realise reusable, scalable and extensible digital twins capable of representing humans in a manufacturing context

Thanks to Industry 4.0, solutions that support monitoring, simulation, optimization and decision making in manufacturing systems are growing exponentially. Most of these solutions rely on the concept of Digital Twins (DTs). NASA introduced the concept of Digital Twin in 2012 as "an integrated multi-physics, multi-scale, probabilistic simulation of a flying vehicle or system" [1]. Since then, this concept has evolved and been adopted in various fields. Thanks to the technologies introduced by the Industry 4.0 paradigm, DTs have gained tremendous importance in the manufacturing industry. DTs have been successfully used for mirroring and simulation of industrial environments, predictive maintenance, virtual commissioning, anomaly detection and product life cycle optimization.

More recently, with the advent of the Industry 5.0 paradigm and the reaffirmation of the crucial role of the worker in production systems, DTs must also enable the representation of humans. The Human Digital Twin (HDT) will be a foundational technology to facilitate the integration of human workers in an Industry 4.0 environment, enabling communication, data aggregation, simulation and planning. In the last 5 years, applications for HDTs have emerged in the manufacturing industry, mainly involving worker monitoring, production planning and control, human-robot collaboration, and adaptive automation. However, the implemented solutions are mostly application-oriented and not reusable.

Currently, there is no solution to support the creation of HDTs, forcing industrial solution architects to resort to ad-hoc implementations and models. On one hand, existing commercial platforms such as Predix or Watson IoT can collect data from a large number of devices and machines but are expensive and require the right support to be integrated in existent production systems. On the other hand, open-source solutions such as Ditto or RAMI AAS, although very interesting for both industry and research, are machine-centric and their information models do not allow to properly model humans.

In STAR, thanks to the close collaboration with the partners of the consortium, the Sustainable Production System Lab of SUPSI is realising a platform to support the development of DTs. The platform enables the representation of humans and, contextual entities in the factory, and their interactions. The resulting platform has been obtained thanks to: (a) a specific information model derived by the metamodel proposed in [2]; (b) a modular infrastructure to ease the instantiation of HDTs and their components [3].

The STAR HDT can be considered as a single source of truth and a central access point to factory entities' data. The HDT embeds the digital representations of workers, which are seamlessly integrated with the DT of the production system and can be used by AI-based modules for a further operation (e.g., prediction tasks). The first prototype of the platform has been released and will be used to support different applications within the project, including workers’ fatigue estimation, AGV path planning, worker training paths definition. The prototype exploits the MQTT protocol to manage data flows from machines and devices to the HDT, and it provides an orchestration module to help the users in defining and managing their own HDT. Users can manage entities (e.g., workers, machines, sensors) and attach AI-based modules to the HDT by means of this component. The platform provides also a dedicated component to manage data flows history, serving collected data as time series for further analytical and monitoring tasks.

By: Niko Bonomi, Andrea Bettoni, Vincenzo Cutrona, Giuseppe Landolfi, Elias Montini and Paolo Pedrazzoli, SUPSI, SPS lab

[1] E. Glaessgen and D. Stargel, "The digital twin paradigm for future NASA and US Air Force vehicles," in Paper for the 53rd Structures, Structural Dynamics, and Materials Conference: Special Session on the Digital Twin, 2012.

[2] Montini, E., Bettoni, A., Ciavotta, M., Carpanzano, E., & Pedrazzoli, P. (2021). A meta-model for modular composition of tailored human digital twins in production. Procedia CIRP, 104, 689-695. DOI: 10.1016/j.procir.2021.11.116

[3] Montini, E., Bonomi, N., Daniele F., Bettoni, A., Pedrazzoli, P., Carpanzano, E., Rocco, P. (2021). The Human-Digital Twin in the manufacturing industry: current perspectives and a glimpse of future. Now publishers -Boston – Delft. DOI:10.1561/9781680838770.ch7