A novel integrated framework for deep-sea submersible search and rescue operations based on trajectory prediction and multi-criteria optimization

Critical safety questions have been recently raised by ocean tourism, particularly following submersible disasters. In the realm of lost submersible rescue, accurate prediction and localization of their position is key due to the complex nature of our marine environments. In this paper, we outlines an integrated framework based on three models, including a physics-based approach for position prediction that takes into account ocean currents, density, and depth context, an Analytic Hierarchy Process (AHP) and Entropy Weight Method (EWM) evaluation system for optimizing the use of rescue equipment, and a Genetic Algorithm (GA)-based probability distribution paradigm for designing search patterns in line with expected areas of interest. The framework also generalizes to cooperative AUV networks for multi-target scenarios. It represents an accuracy of 2% under ± 10% parameter perturbations for the position prediction model. The obtained final ranking of AHP-EWM matrix method revealed that AUVs could be the best primary search equipment followed by ROVs and sonar. In case of the spiral search strategy based on genetic algorithm, it is found that the rates of detection of self-organizing drones are found to be around 85% after 6 hours of its deployment. Such an integrated method composes a full potential framework for submersible rescue operations, demonstrating high reliability across multiple marine contexts, while also presenting practical protocols to adhere to during emergency response functions.