Effective cluster scheduling scheme using local gravitation method for wireless sensor networks

Abstract

Ensuring a certain clustering of distributed sensor nodes is significant for energy efficiency of wireless sensor networks (WSNs). Recently, several energy efficient artificial intelligence based-methods that provide cluster scheduling schemes have been proposed. However, there are few methods providing balanced energy exhausting of the number of nodes in clusters. In this paper, we present a novel cluster scheduling scheme for WSNs using local gravitational method to adaptively determine the number of time-varying clusters based on a sequence of vectors. The main idea in this method is to be inspired by the physical features of gravity force among the mass units. These features are considered to be units of mass of nodes in vector space. It is based on the fact that with clustering, the highest gravitational force on moving the units of mass is applied by the area with feature vectors that are high-density. Cooperative and distributed clustering is formed and the number of clusters is found by sharing the gravitational force between neighboring nodes through a distribution-optimization scheme. The results of this study have been compared with the existing methods in various scenarios and with many performance criteria through the Matlab programming. Especially in dense networks, the proposed algorithm has increased the number of clusters and ensured the clustering balance as much as possible. The proposed method has scaled down the energy consumption, and enhanced the network lifetime.