传感器网络中基于粗糙数据相关的反馈型成簇算法

对于大规模的传感器网络而言，通常采用基于簇的分层路由策略。针对传感器网络中数据的小波压缩，提出了一种基于粗糙数据相关的反馈型成簇算法。该算法首先根据节点数据间的粗糙相关度，形成数据相关性较好的簇结构，然后通过比较Sink反馈的部分小波重构数据与其相应真实数据，进一步优化簇结构。理论分析和仿真实验表明，这种成簇机制使簇内数据的小波压缩具有误差小、压缩比大的优点，且能防止小波压缩中数据淹没现象的产生。

Feed-back clustering algorithm based on coarse data correlation in sensor networks

Large-scale sensor networks usually adopt hierarchical routing strategy.Aiming at wavelet data compression,a Feed-Back clustering algorithm based on Coarse data Correlation(FBCC) in sensor network was proposed.The clusters whose data have better correlations were formed by employing the coarse correlation among sensing data,then some reconstructed data from Sink were compared with the real data in the corresponding sensors and the optimized clusters were obtained.Theoretic analysis and simulation experiments show the proposed algorithm can make the wavelet data compression in sensor networks more efficient,such as smaller error,higher compression ratio.Moreover,it can prevent the data submergence in wavelet data compression.