A real-time approach for smart building operations prediction using rule-based complex event processing and SPARQL query.

Due to the intelligent, adaptive nature of various operations and their ability to provide maximum comfort to the occupants residing in them, smart buildings are becoming a pioneering area of research. Since these architectures leverage the Internet of Things (IoT), there is a need for monitoring different operations (occupancy, humidity etc.) to provide sustainable comfort to the occupants. This work proposes a novel approach for intelligent building operations monitoring using rule-based complex event processing (CEP) and query-based approaches for dynamically monitoring the operations. Siddhi is a CEP engine designed for handling multiple sources of events data in real time and processing it according to predefined rules using a decision tree. Since datastream is dynamic in nature, to keep track of different operations, we have converted the IoT data into an RDF (Resource Description Framework) dataset. The RDF dataset is ingested to Apache Kafka for streaming and for stored data we have used GraphDB tool that extracts information with the help of SPARQL query. Additionally, a large number of events are passed through CEP for correlating and analyzing events in real time, and we measure the performance of CEP engines through SPARQL for enhancing decision support. The proposed approach is evaluated by deploying rules to events through the CEP engine and assessing their processing efficiency in terms of time. Subsequently, a risk estimation scenario is proposed to generate alerts for end users if any smart building operations need immediate attention. However, the proposed work has limitations regarding adaptive and standard parameters, which may cause the rules to fail in adapting to changing data streams. [ABSTRACT FROM AUTHOR]