@article{M0EF76256,
title = "Reinforcement Learning-Based Counter Measure Tactics to Avoid Torpedo Threat",
journal = "The Journal of Korean Institute of Communications and Information Sciences",
year = "2024",
issn = "1226-4717",
doi = "10.7840/kics.2024.49.3.333",
author = "Jaehyun Chung, Gyu Seon Kim, Soohyun Park, Joongheon Kim, Wonhyuk Yun",
keywords = "MDP, Deep Reinforcement Learning, Torpedo Counter Measure Strategies",
abstract = "Recently, research on integrated automation technology for intelligent mission support systems for submarines is being actively conducted in order to maximize combat performance based on future cutting-edge technologies in the submarine system. In particular, advanced technology for submarine torpedo counter measure tactics studies counter measure tactics according to changes in the performance of enemy torpedoes and decoy aircraft.
In general, the speed of a torpedo is much faster than that of a submarine, so quick and accurate calculations are necessary because it is an emergency situation. Accordingly, in this paper, we apply reinforcement learning, one of the machine learning methods, to calculate enemy torpedo counter measure tactics that can respond to any situation and propose a torpedo counter measure tactic algorithm that can select an evasive course for the ship. We apply a reinforcement learning algorithm that can adapt to various variables that may occur in the maritime environment, and through this, we evaluate the excellence of the algorithm and present its applicability to torpedo counter measure tactics in an actual maritime environment."
}