Optimizing Cluster Head Selection with Deep Learning-Based Memory Model (DLMM) for Wireless Sensor Networks

Wireless Sensor Networks (WSNs) play an essential role in a variety of applications due to their capacity for data sensing and transmission. Nonetheless, the finite battery power of sensor nodes poses significant challenges to the longevity of these networks. Traditional routing strategies, which frequently rely on multi-hop transmissions and the creation of clusters, can result in high-energy consumption, particularly for Cluster Heads (CHs) responsible for data aggregation and transmission. This study tackles this issue by employing Deep Learning-Based Memory Model (DLMM) to optimize routing and CH selection for improved energy efficiency. By incorporating a mobile sink that travels along a linear trajectory, the approach minimizes energy expenditure by limiting cluster formation and favouring single-hop transmissions. The method strategically selects CHs based on the nodes’ residual energy levels, thereby prolonging network life. Experimental findings reveal that this strategy can reduce energy consumption by as much as 22.98% in comparison to traditional multi-hop data transmission with circular path sink movements, ultimately enhancing network longevity by 39.05%. The performance assessment, conducted on a 100-node network with varying sink speeds, yielded an energy efficiency improvement of 16.68% over conventional models.