Performance study of zeroth-, first- and second-order data abstraction and reformation algorithms for wireless sensor networks

An energy-efficient data reporting approach in wireless sensor networks is for each sensor node to report only the selected sensed data to the base station, yet the base station can accurately reconstruct the important information later for decision making. Such a process is known as data abstraction and reformation (DAR). The efficiency of a DAR algorithm can then be defined as the accuracy of data reformation against the percentage of the abstracted data over the full data. This paper studies the zeroth-, first- and second-order DAR algorithms and compares their performances in terms of efficiency and other application-oriented requirements.