Context-Aware Scheduling of Distributed DL-Reasoning Tasks in Wireless Sensor Networks

Abstract

Wireless sensor networks (WSNs) provide a means to acquire lots of raw data from vast amounts of easy-to-deploy sensors. Ontologies facilitate structuring data into information and support automatic inference mechanisms. The combination of wireless sensor networks and ontologies can bring significant added value to intelligently process the raw data into meaningful information. In an ontology-based system, this process is referred to as description logics (DL) reasoning. However, the sensors might not be able to execute the reasoning process locally because of resource constraints. Additionally, the usage of the radio interface consumes a lot of power. Therefore, a balance has to be found between local processing and transmission towards the more powerful nodes. In this paper, we present our collaboration platform to bring together wireless sensor networks and distributed ontology A-Box reasoning. This platform should support the adoption of ontology-based methodologies and DL-Reasoning in a distributed setting. We detail a number of algorithms to optimise both bandwidth utilisation and power consumption. These algorithms have been evaluated on a real-life wireless sensor and mesh network test bed, namely WiLab.t. The results show that significant savings of up to 92% in terms of bandwidth utilisation can result from our approach.