Deep learning-based monitoring and localization in autonomous underwater vehicle networks for aquatic ecosystem analysis

Monitoring the aquatic ecosystem is crucial for effective communication and issuing alerts regarding underwater environmental conditions. None of the existing studies have accounted for varying environmental conditions in underwater ecosystem monitoring and communication systems. In this work, sensor data from the Autonomous Underwater Vehicular Network are tuned using a Fuzzy Custom Convex-based Rule to monitor the aquatic ecosystem effectively. Underwater photos and sensor data are pre-processed, features are extracted, correlations are examined, sensor nodes and AUVs are initialized and grouped, cluster heads are chosen, and models are trained to precisely identify species and detect pollution. Finally, pollution is detected using the DSRCN method. The Disaster Management Model is also trained in the same way as the PDM. The DMM attained an accuracy of 98.91%. Thus, the control stations are based on the SMM, PDM, and DMM outputs. Hence, environmental protection is done based on the alert, and the resource is managed accordingly.