LSDIS > Projects > Sensor Map

Semantic Reconciliation with Disparate Sensor Meta-Data for Automatic Publication

PI: Prashant Doshi

A Microsoft SensorMap 2007 RFP award

Project Summary

In both computer science and information science, ontology is a data model that represents a set of concepts within a domain and the relationships between those concepts. It is used to reason about the objects within that domain. Ontologies are used in artificial intelligence, the semantic web, software engineering and information architecture as a form of knowledge representation about the world or some part of it. As ontologies become the preferred ways for storing data, sensor data providers are likely to develop detailed ontologies for their sensor data descriptions. However, currently envisioned and realized frameworks for publication of sensor data, such as the SenseWeb, do not provide a way to utilize provider-defined ontological representations of the sensor data. Instead, they require the provider to undertake potentially tedious and complex ways of registering their sensor feed data with them. Our research considers the provider-defined data models and automatically identifies and semantically aligns relevant concepts from the provider-defined data models with those of the publisher's data models. In addition, we are investigating methods by which the provider data models may be appropriately merged into the publisher's data models, thereby transforming the, possibly minimal, sensor types of the publisher into richer explicit data. This approach will not only alleviate the burden on the data provider by allowing reuse of existing data representations, it will also reduce the burden on the data publisher by avoiding the need to develop detailed data models of the different sensor types. Furthermore, the richer ontologies that result may be used to facilitate refined queries of the sensor data and new combinations with other existing data feeds. The data of third-party sensor feeds will be obtained and primarily used for evaluation. Part of this study includes a proposal to the University of Georgia's campus transit board to set-up a wireless sensor network for tracking the campus bus shuttles with plans for publishing the sensor data on the SensorMap portal. This research is significant because it represents a major step toward automating the publication of sensor data feeds with minimal human effort involved.