Energy management of wireless sensor networks based on multi-layer perceptrons

Efficient energy management is an essential requirement in the design of Wireless Sensor Networks (WSNs). These networks are composed of sensor nodes, which may rely on batteries of limited energy capacity. In addition, many wireless sensor networks are deployed in environments, which are hard to access in order to replace or recharge these batteries. Therefore, maximizing lifetime of such networks is of a paramount importance. A considerable previous research addressed this issue from different angles. One of which is by proposing intelligent models that aim to reduce the number of selected sensors that report environmental measurements. Hence, achieve high energy-efficiency while maintaining a desired level of accuracy in predicting the reported measurement and decision outcomes. The main contribution of this paper is proposing an intelligent model for efficient energy management in WSNs by using the Multi-Layer Perceptrons (MLP) neural network as a classification algorithm. In order to evaluate the performance of the proposed model, several wireless sensor network simulation scenarios based on Ionosphere, Forest CoverType and Sensor Discrimination datasets were conducted. Results show a significant improvement in accuracy level of MLP's prediction compared to Naïve Bayes.