Improving lifetime in wireless sensor networks using neural data prediction

The acceptance of wireless sensor networks (WSN) has increased greatly due to their comprehensive capabilities. Since WSNs are generally battery-powered networks, reducing energy consumption is critical to improve their lifetime and, in turn, their performance and reliability. Recently, smart processing, especially neural networks, has been employed to efficiently manage the power consumed by Wireless Sensor Networks (WSN). Data driven approaches and, in particular, data reduction schemes can reduce the energy spent for communication by judicious selection of the time in which specific sensors of the network are interrogated. In this paper, a multi-layer perceptron (MLP) is used to decide on the data samples required. To justify the usefulness of our idea, we conduct an experiment for effective monitoring of environmental conditions. Results show that our method reduces the number of required samples while not menacing the accuracy needed for practical purposes.