A quality-guaranteed and energy-efficient query processing algorithm for sensor networks

Query processing methods have been studied extensively in traditional database systems. But few of them can be directly applied into sensor database systems due to the characteristics of sensor networks: decentralized nature of sensor networks, limited computational power, imperfect information recorded, and energy scarcity of individual sensor nodes. In this paper, we propose a quality-guaranteed and energy-efficient algorithm (QGEE) for sensor database systems. We employ an in-network query processing method to task sensor networks through declarative queries. Given a query, our QGEE will adaptively form an optimal query plan in terms of energy efficiency and query quality. The goal of our approach is to reduce interference coming from measurements with extreme errors and to minimize energy consumption by providing service that is considerably necessary and sufficient for the requirement of applications. Moreover, we employ probabilistic method to formulate the distribution of imperfect information sources in terms of probability distribution function (PDF), and acquire probabilistic query answers on uncertain data. The probability to an answer allows users to place appropriate confidence in it. Simulation results demonstrate that our algorithm can reduce resource usage and supply quality satisfied query answers to users