Abstract: Due to the inherent uncertainty of sensing data， the processing and optimization techniques for probabilistic Skyline(PS)in wireless sensor networks(WSNs)were investigated. It has been proved that PS was not decomposable after analyzing its properties， so innetwork aggregation techniques could not be used directly to improve the performance. Therefore， a filterbased probabilistic Skyline query processing algorithm in WSNs(FPSP)was proposed to evaluate the PS query in WSNs. The sensing data were divided into candidate data(CD)， relevant data(RD)， and irrelevant data(ID)by the proposed FPSP. The ID in each sensor node could be filtered directly so as to reduce data transmission cost， since PS result could be correctly obtained only according to CD and RD on the base station. The experimental results showed that most of the unnecessary data can be effectively filtered and the lifetime of WSNs can be greatly prolonged by the proposed FPSP algorithm.

Key words: uncertain data, wireless sensor network(WSN), probabilistic Skyline, query processing, data filtering