Lightweight ontology in IoT architecture

The role of ontology in IoT systems is to support the decision making at operative or tactical levels and to increase flexibility and productivity of IoT agents. As the agents have only a limited size memory capacity, the IoT devices can store only lightweight ontology databases. The paper proposes an architecture model and an operational model to implement reduced ontologies in a distributed ontology-server and IoT device framework. One of the key elements in the implementation of reduced ontology is the selection of significant items to be included in the local ontology databases. Based on our comparison tests, the greedy nearest neighbor selection method is a good alternative method for the ontology reduction process.