An integrative framework for cooperative production resources in smart manufacturing

Under the push of Industry 4.0 paradigm modern manufacturing companies are dealing with a significant digital transition, with the aim to better address the challenges posed by the growing complexity of globalized businesses (Hermann, Pentek, & Otto, Design principles for industrie 4.0 scenarios, 2016). One basic principle of this paradigm is that products, machines, systems and business are always connected to create an intelligent network along the entire factory’s value chain. According to this vision, manufacturing resources are being transformed from monolithic entities into distributed components, which are loosely coupled and autonomous but nevertheless provided of the networking and connectivity capabilities enabled by the increasingly widespread Industrial Internet of Things technology. Under these conditions, they become capable of working together in a reliable and predictable manner, collaborating among themselves in a highly efficient way. Such a mechanism of synergistic collaboration is crucial for the correct evolution of any organization ranging from a multi-cellular organism to a complex modern manufacturing system (Moghaddam & Nof, 2017). Specifically of the last scenario, which is the field of our study, collaboration enables involved resources to exchange relevant information about the evolution of their context. These information can be in turn elaborated to make some decisions, and trigger some actions. In this way connected resources can modify their structure and configuration in response to specific business or operational variations (Alexopoulos, Makris, Xanthakis, Sipsas, & Chryssolouris, 2016). Such a model of “social” and context-aware resources can contribute to the realization of a highly flexible, robust and responsive manufacturing system, which is an objective particularly relevant in the modern factories, as its inclusion in the scope of the priority research lines for the H2020 three-year period 2018-2020 can demonstrate (EFFRA, 2016). Interesting examples of these resources are self-organized logistics which can react to unexpected changes occurred in production or machines capable to predict failures on the basis of the contextual information and then trigger adjustments processes autonomously. This vision of collaborative and cooperative resources can be realized with the support of several studies in various fields ranging from information and communication technologies to artificial intelligence. An update state of the art highlights significant recent achievements that have been making these resources more intelligent and closer to the user needs. However, we are still far from an overall implementation of the vision, which is hindered by three major issues. The first one is the limited capability of a large part of the resources distributed within the shop floor to automatically interpret the exchanged information in a meaningful manner (semantic interoperability) (Atzori, Iera, & Morabito, 2010). This issue is mainly due to the high heterogeneity of data model formats adopted by the different resources used within the shop floor (Modoni, Doukas, Terkaj, Sacco, & Mourtzis, 2016). Another open issue is the lack of efficient methods to fully virtualize the physical resources (Rosen, von Wichert, Lo, & Bettenhausen, 2015), since only pairing physical resource with its digital counterpart that abstracts the complexity of the real world, it is possible to augment communication and collaboration capabilities of the physical component. The third issue is a side effect of the ongoing technological ICT evolutions affecting all the manufacturing companies and consists in the continuous growth of the number of threats and vulnerabilities, which can both jeopardize the cybersecurity of the overall manufacturing system (Wells, Camelio, Williams, & White, 2014). For this reason, aspects related with cyber-security should be considered at the early stage of the design of any ICT solution, in order to prevent potential threats and vulnerabilities. All three of the above mentioned open issues have been addressed in this research work with the aim to explore and identify a precise, secure and efficient model of collaboration among the production resources distributed within the shop floor. This document illustrates main outcomes of the research, focusing mainly on the Virtual Integrative Manufacturing Framework for resources Interaction (VICKI), a potential reference architecture for a middleware application enabling semantic-based cooperation among manufacturing resources. Specifically, this framework provides a technological and service-oriented infrastructure offering an event-driven mechanism that dynamically propagates the changing factors to the interested devices. The proposed system supports the coexistence and combination of physical components and their virtual counterparts in a network of interacting collaborative elements in constant connection, thus allowing to bring back the manufacturing system to a cooperative Cyber-physical Production System (CPPS) (Monostori, 2014). Within this network, the information coming from the productive chain can be promptly and seamlessly shared, distributed and understood by any actor operating in such a context. In order to overcome the problem of the limited interoperability among the connected resources, the framework leverages a common data model based on the Semantic Web technologies (SWT) (Berners-Lee, Hendler, & Lassila, 2001). The model provides a shared understanding on the vocabulary adopted by the distributed resources during their knowledge exchange. In this way, this model allows to integrate heterogeneous data streams into a coherent semantically enriched scheme that represents the evolution of the factory objects, their context and their smart reactions to all kind of situations. The semantic model is also machine-interpretable and re-usable. In addition to modeling, the virtualization of the overall manufacturing system is empowered by the adoption of an agent-based modeling, which contributes to hide and abstract the control functions complexity of the cooperating entities, thus providing the foundations to achieve a flexible and reconfigurable system. Finally, in order to mitigate the risk of internal and external attacks against the proposed infrastructure, it is explored the potential of a strategy based on the analysis and assessment of the manufacturing systems cyber-security aspects integrated into the context of the organization’s business model. To test and validate the proposed framework, a demonstration scenarios has been identified, which are thought to represent different significant case studies of the factory’s life cycle. To prove the correctness of the approach, the validation of an instance of the framework is carried out within a real case study. Moreover, as for data intensive systems such as the manufacturing system, the quality of service (QoS) requirements in terms of latency, efficiency, and scalability are stringent, an evaluation of these requirements is needed in a real case study by means of a defined benchmark, thus showing the impact of the data storage, of the connected resources and of their requests