A novel model for efficient cluster head selection in mobile WSNs using residual energy and neural networks

Abstract

Wireless Sensor Networks (WSNs) are essential for monitoring operational environments. Efficient selection of cluster heads is crucial in minimizing energy consumption. This study introduces an innovative model for determining the positions of cluster heads in a WSN, using the residual energy of each sensor node as a selection metric. The sensor incorporates a neural network model that has been trained using different significant models to determine the best location for the cluster head. Our model is compared with the Low Energy Adaptive Clustering Hierarchy (LEACH) algorithm to showcase its effectiveness, particularly in mobile node scenarios. We employ a feed-forward multilayer network, specifically Multilayer Perceptrons (MLPs), as the neural network technique. The proposed method comprises two stages: the training stage, where the neural network is trained with appropriate models, and the execution stage, where the trained network suggests the cluster head's location based on the sensor's conditions. The results demonstrate that the proposed method can accurately identify suitable locations for cluster heads in a sensor network and is capable of adapting to new models despite environmental changes and variations in input configurations.