Abstract : In wireless Sensor Networks (WSNs), energy optimization and efficient routing are NP-Hard problems. Organizing sensor nodes into small clusters with a local coordinator for the collection, aggregation, and transmission of data can reduce energy consumption and prolong the lifetime of WSNs. However, the uniform and balanced clustering of nodes is a challenging task in designing an energy-efficient clustering and routing protocol. One of the simplest yet efficient clustering algorithms that can be employed in WSNs is the K-means algorithm. The paper presents a detailed discussion on the implementation of the K-means algorithm, limitations, and solutions that can be adopted to enhance its performance. The primary objective is to provide a detailed insight into the various approaches employed in WSNs for addressing the limitations of the K-means algorithm. In addition to it, different metrics for evaluating the clustering efficiency of an algorithm are discussed in detail. A tabular review of the characteristics of various modified algorithms, which combine K-means and probabilistic, deterministic, and metaheuristic techniques, is also presented at the end of the work.

Keyword : Approaches, clustering, K-means algorithm, limitations, Wireless sensor networks.