Query processing in wireless sensor networks

Wireless sensor networks are made of autonomous devices, called nodes, that can collect, store, process and share data with other nodes. Users can issue queries over such sensor networks to retrieve data from nodes in applications such as environmental monitoring and surveillance. Query processing is an essential research problem in sensor networks as the usefulness of this new technology depends on its capability to gather and provide data efficiently and effectively when required. In this thesis we investigate query processing techniques for queries over historical sensor observations in a peer-to-peer sensor network. We propose novel techniques for in-network processing of three types of queries: spatial range queries, both exact and approximate and join queries. Due to the limited power supply of the sensor nodes, one of the main challenges of sensor network research is extending the lifetime of the sensor network, and this is the main optimization goal for our techniques. We show, both analytically and through extensive experiments, that our techniques reduce significantly the energy cost of query processing compared to previously existing techniques.