A Submersible Position Prediction and Search-and-Rescue Path Planning Model Based on the Bidirectional Long Short-Term Memory Network

The navigation safety of submersibles is becoming an important issue in this field. In response to submersibles encountering emergencies, we innovatively propose a position prediction model based on the excellent adaptability of the Bidirectional Long Short-Term Memory (BiLSTM) Network to time series, combined with the six-degree-of-freedom motion model of the submersible. This model can predict the position distribution of the submersible within a certain period after losing propulsion power with relative accuracy, as shown in our simulation result of the Ionian Sea. As an application of this technology, we further leverage this position prediction model to introduce a search and rescue path planning model using Monte Carlo Simulation and Genetic Algorithm (GA), providing potential path planning schemes for rescue vessels. This study offers a new perspective and methodology for applying neural network technology to position prediction, serving as a reference for future research.