A wavelet-based sampling algorithm for wireless sensor networks applications

This work proposes and evaluates a sampling algorithm based on wavelet transforms with Coiflets basis to reduce the data sensed in wireless sensor networks applications. The Coiflets basis is more computationally efficient when data are smooth, which means that, data are well approximated by a polynomial function. As expected, this algorithm reduces the data traffic in wireless sensor network and, consequently, decreases the energy consumption and the delay to delivery the sensed information. The main contribution of this algorithm is the capability to detect some event by adjusting the sampling dynamically. In order to evaluate the algorithm, we compare it with a static sampling strategy considering a real sensing data where an external event is simulated. The results reveal the efficiency of the proposed method by reducing the data without loosing its representativeness, including when some event occurs. This algorithm can be very useful to design energy-efficient and time-constrained sensor networks when it is necessary to detect some event.