一种基于线性时间概率计数算法的数据聚集技术

无线传感器网络中，通过数据聚集操作在中间节点预先对数据进行处理，可去除大量冗余，减少数据传输，实现节能。针对多路径路由下数据聚集操作的重复计数问题，研究对副本不敏感的概要结构并优化某些特性，在线性时间概率计数算法的数学模型基础上提出一种新的数据聚集技术FA( Fan Aggregation)技术，实现高能效的数据聚集。理论分析和仿真实验均表明， FA技术相较于FM( Flajolet Martin)技术和LC( Linear Counting)技术在存储空间和准确率上均有更好的性能体现。

A Kind of Data Aggregation Technology Based on Linear Time Probabilistic Counting Algorithm

Using data aggregation technique to pre-processing the data on intermediate nodes in the wireless sensor networks can remove a lot of redundant information,reduce data transmission and save energy. But sensor readings will be over-counted through multipath aggregation computing. To solve this problem,a new data aggregation tech-nique named FA ( Fan Aggregation ) is proposed. Basing on the mathematical model of linear time probabilistic counting algorithm,FA optimizes the specific properties of the duplicate-insensitive synopsis. Theory analysis and simulation results show that FA technique performances better both in storage space and accuracy than FM( Flajolet Martin) and LC( Linear Counting) techniques.