A Confidence Interval Banding Taboo-Level of Data Aggregation Algorithm for Wireless Sensor Networks

Data sending model in cluster has great impact on life of wireless sensor networks. In this paper, the data sending models were researched and a novel CIBT (Confidence Interval Banding Taboo-level) of data aggregation in cluster was formally described, designed, implemented and evaluated. The algorithm includes two stages. In first stage, the sampling data is used to set confidence interval to compare with total model. In second stage, taboo algorithm is adopted to avoid tandem duplication sampling. The performance of CIBT algorithm is evaluated from different aspects. Simulation results show that CIBT can better handle the energy efficiently and satisfy the desired accuracy given by different users. Comparing with similar methods, CIBT can better prolong network life.