IMPROVING LOCALIZATION ACCURACY IN WIRELESS SENSOR NETWORKS BY HYBRID ADAPTIVE WATERWHEEL OPTIMIZATION WITH GENETIC ALGORITHM AND SIMULATED ANNEALING

Abstract

Accurate node localization is critical for the effective operation of Wireless Sensor Networks (WSNs) in applications ranging from environmental monitoring and industrial automation to healthcare and disaster response. However, existing localization techniques often lack scalability, are sensitive to parameters, converge prematurely to local minima, and are susceptible to rogue nodes in dynamic, resource-constrained systems. This paper proposes an enhanced Secure Adaptive Binary Waterwheel Plant Node Localization (SABWP-NL) algorithm that introduces three key innovations: dynamic adaptive parameter tuning based on population diversity, a staged hybridization of SABWP with a Genetic Algorithm (GA) for global diversity injection and Simulated Annealing (SA) for local refinement, and multi-population strategies with restart mechanisms. Simulation results over 30 independent runs demonstrate that SABWP-NL significantly outperforms algorithms such as AO, ROA, BWP, and SABWP. The proposed approach achieves up to 85.88% improvement in the Number of Localized Nodes (NL) and reduces Localization Error (LE) by up to 48.95% across varying anchor node densities. The significance of these improvements is statistically validated with the Wilcoxon rank-sum test $(p < 0.05)$ and non-overlapping 95% confidence intervals. The findings show that SABWP-NL offers a scalable, energy-efficient, and secure localization solution suitable for large-scale, dynamic WSN deployments in real-world applications.