Research on real-time semantic annotation method for sensor data stream

Abstract

In order to conduct the semantic annotation to the heterogeneous, vast and continuous data flow which is captured from micro-environment monitoring platform, inference fresh or implicit knowledge timely according to a new semantic context, and realize real-time monitoring of the micro-environment monitoring platform, the SASML mapping language and the SDRM algorithm are researched and developed, and the S-SASML mapping language and SDS2R algorithm are designed to translate the original sensor data streams into the format of the RDF data streams of the SOSA/SSN. The thread pool techno-logy is used to implement high concurrent processing and improve the real-time performance of our proposed method. The improved mapping language and algorithm can realize the real-time semantic annotation for the continuous, vast data streams on the micro-environmental monitoring platform. The mapping language and algorithm can not only solve the dynamic pickup data flow semantic annotation problem, but also avoid the overload phenomenon caused by high frequency data streams, so that the proposed method has a stable and efficient processing capacity. It basicly meets the demand of micro-environment monitoring platform, and has some application value.

Keywords

• computer perception
• semantic annotation
• real-time
• sensor data flow
• the ontology of SOSA/SSN
• RDF