Computational Intelligence based Clustering Algorithms for Wireless Sensor Networks: Trends and Possible Solutions

A wireless sensor network (WSN) is a state-of-the-art technology for radio communication. A WSN includes several sensors that are arbitrarily distributed in a particular region to detect and track physical characteristics that are hard for humans to observe, like temperature, humidity, and pressure. Because of the nature of WSNs, many issues may arise, including information routing, power consumption, clustering, and cluster head (CH) selection.  Although there are still some difficulties in the WSN, owing to its versatility and robustness, it has gained considerable attention among scientists and technologists despite the shortcomings. Various protocols were designed to solve these problems. Low energy adaptive clustering hierarchy (LEACH) is one of the significant hierarchical protocols used to reduce energy consumption in WSNs. This article provides an extensive analysis of LEACH-variant clustering protocols for WSNs. Recent research on Machine Learning, Computational Intelligence, and WSNs has highlighted the optimized WSN clustering algorithms. However, the selection of a suitable paradigm for a clustering solution continues an issue owing to the diversity of WSN applications. In this paper, a comprehensive review of suggested optimized clustering alternatives has been conducted and a comparison of these optimized clustering methods has been suggested based on various performance parameters. The centralized clustering approaches based on the Swarm Intelligence paradigm are observed to be more suitable for the applications in which low energy is required, high information delivery rate, or elevated scalability than algorithms that are based on the other paradigms described.