Energy saving is an important issue in wireless sensor networks (WSNs) since the nodes are typically powered by batteries with a limited capacity and the energy is limited. As data communication usually consumes much energy and bandwidth, decreasing energy consumption can be generally achieved by reducing the communication of data, for instance, through data compression. Thus, it is an important research issue that how to decrease energy consumption as well as maximize the lifetime of WSNs and increase the efficiency of data communication through data compression on the WSNs nodes where energy supply, memory space and processing resources are constrained. Considering the basic idea of edge operator in the field of image processing and the characteristic that the data stream collected by the nodes of WSNs is time series data, this paper proposes a nonthreshold-based node level algorithm of data compression over the WSNs. The algorithm, with simple calculating and low complexity, compresses the time series data collected by sensor nodes into many piecewise linear representations by extracting some points named edge-points that can indicate the trends of time series data, and especially dose not require any prior knowledge of the monitored objects as well as any predefined threshold value related to the time series data. The experiments on real public sensor data series show that the proposed algorithm can compress data effectively and reduce the communication of data obviously. Moreover, compared with some other data compression algorithms, the proposed algorithm appears better compression performance, reconstructed error and stability which allows the algorithm being applied to the collected data series with different fluctuation characteristics. Consequently, it can save the energy of wireless communication of senor nodes and prolong the lifetime of WSNs better.
[1]
Wang Lilin.
A Simplified Algorithm of Data Compression for Wireless Sensor Network Node
,
2009
.
[2]
Mo Chen,et al.
Data compression trade-offs in sensor networks
,
2004,
SPIE Optics + Photonics.
[3]
Zhancai Yan.
Time Series Pattern Representation Based on Interpolated Edge Operator
,
2007
.
[4]
Eugene Fink,et al.
Search for Patterns in Compressed Time Series
,
2002,
Int. J. Image Graph..
[5]
Liansheng Tan,et al.
A Balanced Serial K-Means Based Clustering Protocol for Wireless Sensor Networks
,
2008,
2008 4th International Conference on Wireless Communications, Networking and Mobile Computing.
[6]
Francesco Marcelloni,et al.
A Simple Algorithm for Data Compression in Wireless Sensor Networks
,
2008,
IEEE Communications Letters.