A neural data-driven approach to increase Wireless Sensor Networks´ lifetime

Wireless Sensor Networks (WSN) play an important role in functioning of various applications. However, technical difficulties, like shortages in power supply, may eventually narrow down WSN´s application range. Minimization of power supply thus can be an adequate mean of prolonging their lifetime. Most of the components of a sensor, including its radio, can be turned off most of the time without influencing the network functionalities it is responsible for. Computational intelligence and, in particular, data prediction methods, may ensure effective operation of the network by the selection of essential samples. In this paper, we apply a multi-layer perception to select the required samples from simulated and experimental meteorological data. The results show that it leads to a considerable reduction of the number of samples and consequently of the power consumption, still preserving the information content.