RDF Stream Based Complex Event Processing In Semantic Web Of Things

As a new research field,Semantic Web of Things is an extension of the Internet of Things.It is characterized by semantic technology to eliminate data heterogeneity,and can combine rich knowledge for semantic query and reasoning.The sensor network,as the infrastructure of Semantic Web of Things,produces highly dynamic data in real time and continuously.RDF streams and the corresponding RDF stream processing techniques have been proposed in order to express and process the dynamic data well.Most of the current RDF stream processing is to extend the SPARQL query for continuous query processing on RDF streams,but how to analyze and identify the time,causality relation in them on the basis of RDF stream querying,namely make a higher level of processing from the perspective of event is less involved.This paper focuses on this issue.In this paper,the existing RDF stream processing technology and the semantic complex event processing method are expounded.On this basis,the RCEP framework for complex event processing based on RDF stream is proposed.Then,the event ontology used in RCEP is modeled.In the process of modeling,the characteristics of the event pattern are taken into account,and the existing SSN ontology is used in combination,which makes the constructed event ontology well expressed.In the process of defining user interested events,appropriate sensors can be selected based on user's settings and metadata in sensor ontology.In the RDF stream processing,Sparkwave is selected as the processing tool.The RETE network structure in Sparkwave is not optimized,this paper using the filter ability of the sub-graph of the event pattern as an indicator to determine its join sequence in RETE network,in order to reduce the number of connections comparison and memory consumption.In Sparkwave RDF reasoning process,in order to reduce the inference time,background knowledge from the ontology is selectively loaded according to the event pattern.At last,the RCEP modules are designed and implemented according to the methods proposed in this paper.In order to verify the improvement of Sparkwave,the performance of improved Sparkwave and original Sparkwave in throughput and memory usage are compared and analyzed.Experiments show that the improved Sparkwave can be effectively applied to complex event detection.