Query Processing Optimization Through Sample Size and Monitoring Coverage Controlling in Wireless Sensor Networks

Abstract

Query processing has been studied extensively in traditional database systems. But few of existed methods can be directly applied to wireless sensor database systems due to their characteristics, such as decentralized nature, limited computational power, imperfect information recorded, and energy scarcity of individual sensor nodes. In this paper, we extend our previous work: quality-guaranteed and energy-efficient algorithm (QGEE) for wireless sensor database systems. We introduce radius of covering disk from point spread function (PSF) aspect and sample size for query quality and energy consumption control. PSF introduces ambiguity into query answers, since the sensitivity of nodes is nonuniform within monitoring region. Sample size determination refers to the process of determining exactly how many samples should be measured in order that the sampling distribution of estimators meets users' pre-specified target precision. In this paper, we formulate the criteria to determine the optimum radius and sample size according to users' requirements on query answers. Simulation results demonstrate that the impact of sample size and monitoring coverage on query answers in terms of root mean square error (RMSE)