Semantic access to sensor observations through web APIs

Abstract

Sensor networks are often deployed with the purpose of providing data to large-scale information management and GIS systems, or to collect measurements for specific scientific experiments. The benefits of such use are clear and widely accepted. The reuse of observations in low-cost, lightweight, web applications and mashups is a further compelling use case for sensor networks, but requires provision of data through simple mechanisms, readily accessible, that are quick to develop with. To enable the latter while maintaining support for larger applications and, to increase information utility, links to and from other datasets, we propose a domain-driven approach that embodies REST and Linked Data principles using a common semantic framework that underpins a separation of concerns between domain models, sensor observation infrastructure, and Application Programming Interfaces (APIs) while maintaining information consistency. We describe a reusable, reconfigurable, web service that realises this design and can be deployed to provide access to multiple sources of sensor information, including databases and streaming data, with flexible semantic configuration of the API and domain mapping. © 2011 IEEE.