Submerged Sensor Localization Using Cayley-Menger Determinant with Diverse Equations and Security

This thesis introduces a novel approach for underwater sensor localization by employing a static configuration of six surface beacons arranged in a pentagonal formation. This spatial arrangement effectively mitigates singularity issues in the Cayley-Menger determinant which typically arise in co-linear or co-planar configurations to ensure accurate and reliable position estimation. The proposed method incorporates real-time filtering techniques to account for minor sensor displacements caused by ocean currents or sinking to enhance the localization system's robustness. The predictive model utilizes machine learning algorithms such as Linear Regression, Random Forest, and Recurrent Neural Networks (RNNs), is employed to estimate the future positions of submerged sensors while compensating for the dynamic effects of ocean currents. Additionally, this thesis evaluates the acoustic speed using multiple empirical formulas—Mackenzie, Medwin, Chen–Millero, Wilson, Leroy, Coppens, and Del Grosso—each tailored to different environmental conditions, ensuring precise distance measurement and mitigating inaccuracies in position estimation. The effect of three environmental factors has been measured and analyzed in detailed. Furthermore, security concerns related to underwater sensor networks are addressed through the development of jamming detection protocols which will ensure reliable communication despite the challenging nature of underwater acoustic environments. This comprehensive approach provides a scalable, accurate, and efficient solution for underwater sensor localization and monitoring also offering significant advancements in submerged sensor tracking and marine research under dynamic and variable oceanic conditions.