Edge Computing-Based Signal Processing for Optimized-Communication and Anomaly Detection in IoUT Using TM-TCO and HMC-FRL

The detection and monitoring of anomalies in the underwater ecosystem are necessary to mitigate disasters. None of the traditional works concentrated on detecting the underwater anomalies regarding boundaries of the tectonic plates. Therefore, anomaly detection in the underwater ecosystem, particularly in relation to tectonic plate boundaries, is proposed. Using Dynamic Bit Permutation-SHA (DBP-SHA) hashing, Internet of Underwater Things (IoUT) devices are registered in a Multi-Hop Network. Tent Map Termite Colony Optimization (TM-TCO) optimizes pathways, whereas Terahertz Quantum Resistant Cryptographic Algorithm (Te-QRCA) guarantees secure data transfer. Density-Based Spatial Clustering of Applications with Noise (DB-SCAN) balances workload, while edge servers use the Exponential Gaussian Peak (EGP)-Fuzzy model to train local models that assess less sensor data, spot disaster trends, and produce labels. Then, the Hierarchical Meta-Curiosity-based Federated Reinforcement Learning (HMC-FRL) classifier is used for anomaly detection. The trained local model is updated on the cloud server. Thus, the communication is optimized and the anomaly is detected effectively in the IoUT with an accuracy of 98.671%.