Submersible Position Prediction and Identification Based on Hybrid Neural Network Algorithm and Monte Carlo Method

This paper proposes a method for predicting and identifying the position of a submersible using a hybrid neural network algorithm and the Monte Carlo method. Initially, we employ a hybrid neural network algorithm to establish a predictive model for ocean current parameters based on data from the Ionian Sea detection station. Additionally, we analyze the forces acting on the submersible and its motion to construct a physical model for the submersible’s vertical movement. By integrating the effects of these two models, we can predict the three-dimensional trajectory of the submersible. Furthermore, to accurately analyze the potential location of the submersible after an accident, we utilize the Monte Carlo method and infinitesimal method, employing the K-means clustering algorithm to analyze the initial deployment points of the device. Based on the probability distribution, we select appropriate search methods. As indicated by the established probability function model, the probability of the submersible being covered decreases over time. The search results serve as adjustment parameters for the probability function model; the more search results obtained, the higher the probability of locating the submersible is.