Abstract: As the high computational complexity, low compression efficiency and data recovery rate of current data compression methods of Wireless Sensor Network（WSN）, a WSN data compression method based on a cluster head base station separation structure is proposed. Based on the monolayer cluster structure of WSN, the method requests the terminal sensor node to transmit initial data in segments at first. Then the original WSN data compression method is used to compress the data received by the cluster head nodes. Finally, the grey model is adopted for data recovery at the base station. In addition, the optimal model and segment length of the algorithm are given, through the experimental analysis of the data recovery effect of the grey model and the grey Markov chain model. Experimental results show that, compared with the traditional based linear regression method, the proposed method can significantly improve the accuracy of data recovery at higher compression efficiency.

Key words: wireless sensor networks, spatial correlation, data compression, grey mode

摘要： 针对目前无线传感器网络（WSN）数据压缩方法的计算复杂度高、压缩效率和数据恢复准确率较低的情况，提出基于簇头-基站分离式结构的WSN数据压缩方法。该方法在WSN的单层分簇结构的基础上，要求感知节点将采集的原始数据分段发送，采用原有WSN数据压缩方法对簇头节点接收的数据进行空间相关性压缩，在基站采用灰色模型进行数据恢复。另外，通过实验分析灰色模型与灰色马尔可夫链模型对数据的恢复效果，给出算法最优模型与段长。仿真结果表明，提出的方法相比传统线性回归方法在较高压缩效率时可显著提高数据恢复精度。

关键词: 无线传感器网络, 空间相关性, 数据压缩, 灰色模型