D5.2 Enablers for Machine Initiated Semantic IoT Search

by | Jul 7, 2020 | 0 comments

This deliverable concludes the task T5.2. The objectives of this task is to define and develop context-aware, semantic search algorithms for large-scale IoT resources. Motivation for this is a limited functionality of the query language for NGSI-LD, which is selected as metadata format for IoTCrawler. The deliverable exploits the context defined in T5.1 in the form of Topical Domain Ontologies (TDO) to generate queries for machine-initiated search. A context-awareness is enabled via query generation and processing based on the GraphQL language, where the initial language was extended with features for traversing entities relationships, integration of high-level domain semantics, nested entity filtration and others. The context-awareness of GraphQL queries is based on a newly introduced state-based context model, a rule-/pattern-based generator for state transitions, and a mapping mechanism for generating filter conditions based on active context states kept in the metadata repository. The described search functionality is functioning on top of the IoTCrawler’s federated metadata infrastructure integrated with security and privacy-aware mechanisms. Several examples of GraphQL schemas were designed in terms of TDOs to demonstrate the value of the developed search engine for IoT applications targeted for three different domains.

Acknowledgement of other key-contributors: Martin Strohbach (AGT), Patrik Schneider (Siemens), Pedro Gonzalez Gil, Antonio Skarmeta (UMU), Tarek Elsaleh (UniS), Aurora Gonzalez (UMU), Roonak Rezvani (UniS) and Hien Truong (NEC).

Pavel Smirnov


WordPress Appliance - Powered by TurnKey Linux