DCT Compression Scheme for Wireless Sensor Information considering Spatial Correlation
Kohei Natsume, Mai Ohta, Koichi Adachi, Osamu Takyu, Takeo Fujii · 2021
The data compression techniques effectively reduce energy consumption at a sensor in wireless sensor networks (WSNs). To save energy, a simple and high performance compression method should be chosen. In this paper, we propose a discrete cosine transform (DCT) compression technique considering spatial correlation. The proposed method sorts a series of sensor nodes (SNs) to minimize the spatial Euclidean distance between SNs. The computer simulation using real temperature in the room shows the proposed method can improve reconstruction accuracy. When the allowable error is set at 0.5 [degree Celsius], compared with the conventional DCT, the data that achieves allowable error improve more than 10 [%].