Artificial Neural Networks‐Based LEACH Algorithm for Fast and Efficient Cluster Head Selection in Wireless Sensor Networks

• Published in: International journal of communication systems Vol. 38; no. 3
• Main Author: Senturk, Arafat
• Format: Journal Article
• Language: English
• Published: 01.02.2025
• Subjects: artificial neural networks; cluster head selection; LEACH; machine learning; wireless sensor networks
• ISSN: 1074-5351; 1099-1131
• DOI: 10.1002/dac.6127

ABSTRACT Recent improvements in wireless sensor networks (WSN) technology enabled more research on energy efficiency. Limited battery and difficulties in renewing the batteries in a critical application require the efficient use of energy for WSN. Besides energy efficiency, providing fast‐response systems facilitate real‐time applications. Combining machine learning (ML) with the clustering methods that significantly contribute to the energy efficiency of WSN seems to improve the efficiency. In this paper, the low‐energy adaptive clustering hierarchy (LEACH) clustering method for WSN is implemented through a supervised learning method, artificial neural networks (ANN), for cluster head (CH) selection. The power of ANN as a superior classifier is thought to contribute much to the field. The details of designing an ANN are given in detail for the first time in WSN field. In addition, a dataset is prepared via MATLAB to be used for classification or other analysis related to a clustered network. The proposed model provides more than 85% accuracy for CH selection, and it is 83.28% faster than LEACH to determine the CHs. This method produces more efficient solutions in large networks in terms of the time for CH selection. The feasibility of ANN is also shown for the issues related to WSN such as CH selection. ANN determines cluster heads (CH) 85% similar to LEACH with less computational load. ANN determines CHs more than 83% faster than LEACH, which supports real time applications. ANN can be adopted easily to various WSN problems and provides efficient solutions. CH selection can be investigated as a classification problem and can be solved by machine learning algorithms.