Complex events and actions to control cyber-physical systems

  • Authors:
  • Rüdiger Klein;Jingquan Xie;Andrij Usov

  • Affiliations:
  • Fraunhofer IAIS, Sankt Augustin, Germany;Fraunhofer IAIS, Sankt Augustin, Germany;Fraunhofer IAIS, Sankt Augustin, Germany

  • Venue:
  • Proceedings of the 5th ACM international conference on Distributed event-based system
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Cyber-Physical Systems (CPS) are controlled complex technical systems. On one hand they do exist and behave in the physical world with their dedicated physical and technical rules. On the other hand they are controlled according to certain policies -- frequently by a "tandem" of control systems and human operators. Modern cars, aircrafts, power grids, production lines etc. are typical cyber-physical systems. The unprecedented technology development has enabled these systems to play an increasingly central role in modern societies. However, to ensure the reliability and effectiveness of CPS operations is always a challenging task since they are often heterogeneous, large-scale and very complex for conventional software systems. In this paper we propose a novel rule-based approach combining deductive and reactive rules to specify and describe CPS. Complex Event Processing (CEP) and Event-Condition-Action (ECA) rules are used to detect situations of interest in CPS and issuing (complex) control actions, respectively. This allows us to focus on the "what" of CPS control (situation detection, reaction specification) and leaves the "how" to a well defined event and action engine. Whereas events are treated in CPS much the same as in other fields actions are quite different because they happen in the real world. We describe an approach to CPS actions which takes these important particularities into account. Another important issue is the integration of physical models into event and action processing. Temporal aspects play a central role here. Our approach enables us to control CPSs under both normal and exceptional conditions in a flexible and efficient way. In order to illustrate the basic features and expressivities of the proposed approach we consider realistic industrial situations.