Histogram Estimation for Optimal Filter Skyline Query Processing in Wireless Sensor Networks

Abstract

The skyline query processing technique plays an increasingly important role for multicriteria decision making applications in wireless sensor networks. The technique of saving energy to prolong the lifetime of sensor nodes is one of the dominating challenges to resource-constrained wireless sensor networks. In this paper, we propose an energy-efficient skyline query processing algorithm, called the histogram filter based algorithm (HFA), to efficiently retrieve skyline results from a sensor network. First, we use historical data at the base station to construct histograms for further estimating the probability density distributions of the sensor data. Second, the dominance probability of each tuple is computed based on the histograms, and the optimal tuple which has the largest possibility of dominance/filtering capability is obtained using in-network aggregation approach. After that, the base station broadcasts the optimized tuple as the global filter to each sensor node. Then, the tuples which do not satisfy the skyline query semantics are discarded to avoid unnecessary data transmissions. An extensive experimental study demonstrates that the proposed HFA algorithm performs more efficiently than existing algorithms on reducing data transmissions during skyline query processing, which saves the energy and prolongs the lifetime of wireless sensor networks.