Semantic-driven CEP for delivery of information streams in data-intensive monitoring systems

Abstract

Modern control and monitoring systems incorporate large number of heterogeneous devices and are operated by many users with different roles and responsibilities. The information generated by these devices, although preprocessed and filtered, is usually delivered to users regardless of their actual information needs, thus overwhelming cognitive capacities and potentially affecting safety of the system. This paper addresses the problem of personalized information delivery and presents hybrid system capable to deliver information tailored to user needs based on his/her role and responsibilities. The system utilizes Semantic Web technologies (specifically OWL-DL ontologies and SPARQL queries) for configuration purposes and Complex Event Processing (CEP) for run-time analyzing. The combination of mentioned technologies brings two major advantages: (1) the behavior of the system could be easily changed by configuring only underlying ontology;(2) utilization of CEP at run-time makes system event-driven and reactive to frequent changes in the environment. It is expected that proposed approach is able to make monitoring systems personal oriented and thus safer during the operation. The paper presents implemented software tools and illustrates approach with scenarios from building automation domain. Current implementation of the proposed approach suggests feasibility of the solution and prompts directions for future research.