Energy efficiency is one of the main challenges in developing Wireless Sensor Networks (WSNs). Since communication has the largest share in energy consumption, efficient routing is an effective solution to this problem. Hierarchical clustering algorithms are a common approach to routing. This technique splits nodes into groups in order to avoid long-range communication which is delegated to the cluster head (CH). In this paper, we present a new clustering algorithm that selects CHs using the grey wolf optimizer (GWO). GWO is a recent swarm intelligence algorithm based on the behavior of grey wolves that shows impressive characteristics and competitive results. To select CHs, the solutions are rated based on the predicted energy consumption and current residual energy of each node. In order to improve energy efficiency, the proposed protocol uses the same clustering in multiple consecutive rounds. This allows the protocol to save the energy that would be required to reform the clustering. We also present a new dual-hop routing algorithm for CHs that are far from the base station and prove that the presented method ensures minimum and most balanced energy consumption while remaining nodes use single-hop communication. The performance of the protocol is evaluated in several different scenarios and it is shown that the proposed protocol improves network lifetime in comparison to a number of recent similar protocols.

Introduction

Wireless sensor networks (WSNs) are emerging low-cost and versatile solutions that enable controlled monitoring of the environment. They generally consist of a large number of small sensing devices that are capable of data processing and wireless communication. These sensor nodes can be deployed in various environments to implement applications such as habitat monitoring, military surveillance, home and industrial automation, and smart grids [1], [2]. Recent advances in electronic circuit design have made it possible to build lighter, cheaper and more energy efficient sensors. However many research areas including energy efficiency need to be further studied [3]. In many applications, sensor nodes are equipped with a non-rechargeable battery that restricts network lifetime [4]. There are several definitions for lifetime, such as the time until the first node dies or the time that the last node dies or the time until a specific fraction of nodes die [5]. After the death of the first node, the performance of the network will degrade sharply [6]. In [7] and [8], network lifetime is defined in terms of node lifetime, coverage, and connectivity. Although the use of renewable energy sources for sensor nodes are investigated in Energy Harvesting Wireless Sensors Networks (EHWSN) [9], wise use of the available energy is still required for long running WSNs. Most WSNs measure physical parameters such as temperature, humidity or location of objects.