一种能量高效的WSN近似数据收集算法

为降低能耗和最大化网络生存期，论文提出了在一定误差范围内的高效近似数据收集算法。首先利用节点感知数据的时间相关性生成局部估计模型，然后根据节点间估计数据的空间相关性进行分簇，在簇首进行相关性检测，动态调整簇结构，并将簇首的模型参数上传给SINK节点，最后在SINK节点进行全局近似数据收集。仿真结果表明，该算法能充分利用节点数据的时空相关性去除冗余数据，在给定误差界限内能显著降低通信成本。

Energy-efficient Algorithm of Approximate Data Collecting in WSN

An energy-efficient algorithm of approximate data collecting with given error bounds is proposed to reduce the energy con- sumption and maximize the network lifetime in wireless sensor networks. First, the algorithm uses the temporal correlations of sensor read- ings to build these local models at each node. Then the algorithm based on the spatial correlations between estimates data of sensor nodes for clustering. During the data collection procedure, the data correlations are detected in the cluster heads to insure better correlation and adjust the clusters structure, then the model parameters of cluster heads are transmitted to SINK. Finally the algorithm per{orms a global approxi- mate data collection on SINK according to model parameters uploaded by cluster heads. The simulation results show that the proposed algo- rithm can eliminate the data redundancy by using temporal and spatial correlation as much as possible, and notably reduce the communication cost with given error bounds.