Abstract: The presence of spatial correlation in in-cluster data has been exploited to minimize communication costs（and hence，energy costs） incurred during data gathering in a sensor network.Using source coding scheme to compress the correlated information that observed by multiple sensors，quantize the correlated observations independently，and transmit independent bits to a central processor，which is responsible for reconstructing the physical environment with the smallest distortion possible.This paper proposes a DCT-based In-cluster Data Aggregation Scheme（DIDAS），in which the data aggregation makes use of the characteristic of discrete cosine transform.Simulation results show that the proposed scheme performs well in compression scale and distortion degree.

Key words: wireless sensor network, spatial correlation, data aggregation, DCT

摘要： 无线传感器网络中同一簇内采集到的数据具有空间相关性，由此提出了一种基于DCT变换的簇内数据融合方案。在信源编码中，利用空频变换方法，在簇头节点对簇内数据去除空间上的相关性，从而实现数据融合，并使得SINK节点重构恢复的数据与原始数据质量基本相同。经过MATLAB仿真分析，该数据融合方案在压缩比和失真度上均表现出良好的特性。

关键词: 无线传感器网络, 空间相关性, 数据融合, DCT变换