隐蔽是潜艇作战的最大优势。对作战海域潜艇隐蔽性的仿真及预测是其航行的依据和准则。本文假设某海域不同位置各有一艘潜艇和主动声呐舰艇，二者先 “十”字航线机动，再 “Z”字航线机动。收集该海域大量海洋环境数据，采用Bellhop模型计算出舰艇主动声呐的信号余量，进而推算潜艇隐蔽等级作为参考值。引入有序多分类Logistic模型学习“十”字机动航线上的数据，再预测“Z”字机动航线上的潜艇隐蔽性等级。结果表明： Logistic模型计算速度较Bellhop模型快70多倍。在模型训练中，多个变量显著，模型能够很好的拟合训练数据。引入多分类评价指标，模型预测准确率达到0.801，精确率0.871，召回率0.788，F1指标0.827。Logistic模型对潜艇隐蔽性预测快速可靠，可为作战中潜艇隐蔽性提供辅助作用。
Concealment is the biggest advantage of submarine warfare. The simulation and prediction of the concealment of combat waters are the basis and criteria for submarine navigation. This study assumes that there are one submarine and one active sonar vessel at different locations in a certain sea area, and they first maneuver along the "ten" route and then along the "Z" route. Collect a large amount of marine environmental data in the sea area, use the Bellhop model to calculate the signal margin of the ship's active sonar, and then calculate the submarine′s concealment level. Introduce an ordered multi class logistic model to learn data on the "ten" shaped maneuvering route, and then predict the submarine stealth level on the "Z" shaped maneuvering route. The results indicate that the logistic model is 70 more times faster than the Bellhop model in terms of computation speed. In model training, multiple variables are significant, and the model can fit the training data well. By introducing multiple classification evaluation indicators, the model achieved a prediction accuracy of 0.801, an accuracy of 0.871, a recall rate of 0.788, and an F1 index of 0.827. The logistic model is fast and reliable for predicting submarine concealment, and can provide auxiliary functions for submarine concealment in combat.
2025,47(14): 103-111 收稿日期：2024-10-22
DOI：10.3404/j.issn.1672-7649.2025.14.016
分类号：U674.76
基金项目：国家自然科学基金资助项目（72371137）
作者简介：吴穹(1988-),男,博士,讲师/工程师,研究方向为水下声场建模、大规模海洋数据分析与处理、机器学习模型
参考文献：
[1] 王磊, 常书刚. 潜艇噪声与综合降噪技术的应用[J]. 航海技术, 2007(2): 44-48.
WANG L, CHANG S G. Application of submarine noise and comprehensive noise reduction technology[J]. Navigation Technology, 2007(2): 44-48.
[2] 孙雪海, 笪良龙, 李玉阳. 水下三维声场仿真与可视化方法研究[J]. 舰船科学技术, 2015, 37(9): 116-120.
SUN H X, DA L L, LI Y Y. Research on underwater 3D acoustics energy field simulation and visualization method[J]. Ship Science and Technology, 2015, 37(9): 116-120.
[3] 张驰, 马晓川, 李璇, 等. 千岛湖复杂地形条件下传播损失的估计与现场测量[J]. 信号处理, 2017, 33(3): 367-373.
ZHANG C, MA X C, LI X, et al. Estimation and on-site measurement of propagation loss under complex terrain conditions in Qiandao Lake[J]. Signal Processing, 2017, 33(3): 367-373.
[4] 魏士俨, 杨燕明, 阮海林. 跨大洋水声传播建模与绕射效应[J]. 地球物理学报, 2024, 67(8): 3253-3264.
WEI S Y, YANG Y M, RUAN H L. Modeling and diffraction effects of cross ocean acoustic propagation[J]. Chinese Journal of Geophysics, 2024, 67(8): 3253-3264.
[5] HUANG J G, WANG H, HE C B, et al. Underwater acoustic communication and the general performance evaluation criteria[J]. Frontiers of Information Technology & Electronic Engineering, 2018, 19(8): 951-971.
[6] NIU H , REEVES E , GERSTOFT P. Source localization in an ocean waveguide using supervised machine learning[J]. The Journal of the Acoustical Society of America, 2017, 142(3): 1176-1177.
[7] KATSNELSON B, PETNIKOV V, LYNCH J. Fundamentals of shallow water acoustics [M]. New York: Springer, 2012: 425-435.
[8] 蒋廷华, 华孝先, 吴维琴. 水中有限长弹性圆柱的目标强度研究[J]. 声学学报, 1985(2): 85-95.
JIANG Y H, HUA X X, WU W Q. On target strength of an elastic cylinder with finite length in water[J]. Chinese Journal of Acoustics, 1985(2): 85-95.
[9] 赵志允, 孙明太, 任东彦. 吊放声呐主动工作方式探测域建模与仿真[J]. 指挥控制与仿真, 2013, 35(6): 49-53.
ZHAO Z Y, SUN M T, REN D Y. Modeling and simulation of detection domain for active working mode of suspended sonar[J]. Command and Control and Simulation, 2013, 35(6): 49-53.
[10] GJB 6850.8-2009水面舰船系泊和航行试验规程 第8部分: 适航性试验[S].
[11] 吴福初, 徐寅, 蓝天. 单舰巡逻反潜战术行动方法及对潜警戒能力分析[J]. 兵工自动化, 2016, 35(12): 45-48.
WU F C, XU Y, BLUE S. Analysis of single ship patrol anti submarine tactical operation methods and submarine vigilance capability[J]. Ordnance Automation, 2016, 35(12): 45-48.
[12] 陈桂, 刘洋. 基于人工智能的断层自动识别研究进展[J]. 地球物理学进展, 2021, 36(1): 119-131.
CHEN G, LIU Y. Research progress on automatic fault recognition based on artificial intelligence[J]. Advances in Geophysics, 2021, 36(1): 119-131.
