Tauber, Markus

2019-10, Maksuti, Silia, Tauber, Markus, Delsing, Jerker

Cyber-physical production systems are engineered systems that are built from, and depend upon, the seamless integration of computational algorithms and physical components. In order to make these systems interoperable with each other for addressing Industry 4.0 applications a number of service-oriented architecture frameworks are developed. Such frameworks are composed by a number of services, which are inherently dynamic by nature and thus imply the need for self-adaptation. In this paper we propose generic autonomic management as a service and show how it can be integrated in the Arrowhead framework. We propose generic and reusable interfaces for each phase of the autonomic control loop in order to increase the usability of the service for other frameworks and application systems, while reducing the software engineering effort. To show the utility of our approach in the Arrowhead framework we use a climate control application as a representative example.

2019-02, Maksuti, Silia, Schluga, Oliver, Settanni, Giuseppe, Tauber, Markus, Delsing, Jerker

Manufacturing enterprises are constantly exploring new ways to improve their own production processes to address the increasing demand of customized production. However, such enterprises show a low degree of flexibility, which mainly results from the need to configure new production equipment at design and run time. In this paper we propose self-adaptation as an approach to improve data transmission flexibility in Industry 4.0 environments. We implement an autonomic manager using a generic autonomic management framework, which applies the most appropriate data transmission configuration based on security and business process related requirements, such as performance. The experimental evaluation is carried out in a MQTT infrastructure and the results show that using self-adaptation can significantly improve the trade-off between security and performance. We then propose to integrate anomaly detection methods as a solution to support self-adaptation by monitoring and learning the normal behavior of an industrial system and show how this can be used by the generic autonomic management framework.