A Novel Trust-Aware and Energy-Aware Clustering Method That Uses Stochastic Fractal Search in IoT-Enabled Wireless Sensor Networks

Wireless sensor network (WSN) technology is considered to be an integral part of large-scale and efficient deployment of Internet-of-Things (IoT). More specifically, in mission-critical IoT applications, trust in the sensor data is becoming increasingly important. Sensor nodes have limited processing, storage, and communication capabilities, which make them susceptible to attacks and unreliable functioning. However, the limitations in the energy resources of the sensors are a major challenge in maximizing the network’s lifetime. Grouping the sensors into clusters was proposed to address such energy limitations. Many meta-heuristic clustering protocols have been proposed to maximize the network lifetime, which is an NP-hard problem. This problem is more complicated when considering the trust factor. The majority of existing clustering models were built to reduce the energy consumption in the network without considering the energy consumption required to detect untrusted nodes, and thus, it requires extra energy consumption to perform this task. This article proposes a clustering protocol with a trust model that detects the untrusted nodes through energy and data-trust. In addition, the proposed clustering protocol maximizes the network’s lifetime through the good characteristics of stochastic fractal search optimization. Finally, a novel fitness function is introduced to select the cluster-heads among the trusted nodes. The function is based on the following four parameters: 1) the remaining energy of the nodes; 2) the density of the nodes; 3) the distance between each node and the base-station; and 4) the network’s dissipated energy. When forming the clusters, the density of the cluster-heads is considered to balance the load of all of the cluster-heads. The experimental evaluation performed here affirms the efficacy of the proposed protocol in comparison with existing protocols.

I. Introduction

Internet-of-things (IoT) applications have become an integral part of our life. IoT networks are heavily dependent on wired and wireless identifiable devices. Several IoT networks use wireless sensors that are connected to a centralized device (e.g., a router), which is responsible for managing the operations of the sensors and their connections to the Internet. Wireless sensor networks (WSNs) are integrated within IoT deployments when networked to the Internet. A WSN is a group of large number of sensor nodes and a base-station (so-called sink node). The sensor nodes have limited processing, storage, and communication capabilities. WSNs have several applications, such as military operations, environmental monitoring, and disaster management [1]. For monitoring and early-warning IoT-enabled WSNs, the collected information from the sensor nodes must be trustworthy. In such systems, the IoT sensor devices will detect the event of interest and generate an alert packet for warning and early detection of that event, such as the gas leak detection IoT system in [2]. However, there is a chance for these sensor nodes to behave unreliably due to their limited capabilities and the inhospitable physical environments in which these nodes are deployed [3], [4]. To save energy and increase the number of live nodes, one of the most commonly used solutions is clustering, where a group of sensor nodes forms a cluster, and only the cluster-head sensor node is responsible for interacting with the base-station [5]. Clustering is efficient because it can prevent exchanging redundant messages and can maintain the communication bandwidth, stabilize the topology of the network, and reduce the communication overhead.