海洋信道受湍流效应干扰，使得在舰船跳频/扩频体制下通信数据呈现非平稳随机特性，难以在非平稳状态下提取数据的时序信道均衡度特征，降低了数据传输稳定性。为此，提出基于长短期记忆（Long Short-Term Memory，LSTM）的舰船通信大数据异常探测方法。构建舰船通信大数据平台架构中舰船端与岸端数据传输信道模型，计算舰船通信大数据平台舰船端与岸端传输舰船数据量，提取时序信道均衡度作为关键特征，将关键特征输入到LSTM模型中，利用其门控机制处理时序特征，最终实现舰船通信大数据异常探测。实验证明该方法能够深度挖掘微小时序信道均衡度特征，有效检测出舰船通信大数据异常状态，对异常结果进行防范后，舰船大数据通信带宽、吞吐量显著提升，舰船数据传输稳定性提升至98.5%。
The ocean channel is disturbed by the turbulence effect, which makes the communication data in the ship frequency-hopping/spread-spectrum system present non-stationary random characteristics. It is difficult to extract the temporal channel equalization degree characteristics of the data in the non-stationary state, reducing the stability of data transmission. Therefore, an anomaly detection method for ship communication big data based on Long Short-Term Memory(LSTM) is proposed. Construct the data transmission channel model between the ship end and the shore end in the architecture of the ship communication big data platform, calculate the amount of ship data transmitted between the ship end and the shore end of the ship communication big data platform, extract the equalization degree of the temporal channel as the key feature, input the key feature into the LSTM model, use its gating mechanism to process the temporal features, and finally achieve anomaly detection of ship communication big data. Experiments prove that this method can deeply mine the equalization degree characteristics of micro-time series channels, effectively detect the abnormal states of ship communication big data. After preventing the abnormal results, the communication bandwidth and throughput of ship big data have significantly increased, and the stability of ship data transmission has improved to 98.5%.
2025,47(22): 166-170 收稿日期：2025-6-12
DOI：10.3404/j.issn.1672-7649.2025.22.024
分类号：U665.2；TP393
基金项目：辽宁省教育科学十四五规划课题（JG21DB081）
作者简介：李秋（1979 – ），女，硕士，讲师，研究方向为大数据、数据分析和算法等
参考文献：
[1] 吴强. 能量受限约束下的舰船通信网络节点动态协同部署[J]. 舰船科学技术, 2023, 45(4): 171-174.
WU Q. Dynamic cooperative deployment of ship communication network nodes under energy constraints[J]. Ship Science and Technology, 2023, 45(4): 171-174.
[2] 郭昊, 李海滨, 冯姣, 等. 基于大数据处理的船舶数据质量评价方法研究[J]. 计算机仿真, 2022, 39(2): 298-303.
GUO H, LI H B, FENG J, et al. Ship AIS data quality evaluation algorithm based on big data processing[J]. Computer Simulation, 2022, 39(2): 298-303.
[3] 李费旭, 周利, 丁仕风, 等. 基于改进LSTM的船体监测数据异常处理方法[J]. 船舶工程, 2024, 46(7): 90-102+121.
LI F X, ZHOU L, DING S F, et al. Exception handling method for hull monitoring data based on improved LSTM[J]. Ship Engineering, 2024, 46(7): 90-102+121.
[4] 彭耀武, 陈辰, 刘敬贤, 等. 基于BO-GRU和AKDE的船舶异常行为识别[J]. 中国航海, 2024, 47(3): 10-20.
PENG Y W, CHEN C, LIU J X, et al. Identification of abnormal ship behavior based on BO-GRU and AKDE[J]. Navigation of China, 2024, 47(3): 10-20.
[5] 张俊峰, 吴双. 考虑异常数据及船舶行为的在线AIS轨迹压缩算法[J]. 大连工业大学学报, 2024, 43(6): 462-468.
ZHANG J F, WU S. An online AIS trajectory compression algorithm considering abnormal data and ship behaviors[J]. Journal of Dalian Dalian Polytechnic University, 2024, 43(6): 462-468.
[6] 张浩晢, 杨智博, 焦绪国, 等. 基于增强Bi-LSTM的船舶运动模型辨识[J]. 中国舰船研究, 2025, 20(1): 76-84.
ZHANG H Z, YANG Z B, JIAO X G, et al. Ship motion identification model based on enhanced Bi-LSTM[J]. Chinese Journal of Ship Research, 2025, 20(1): 76-84.
[7] 廉清云, 孙伟, 李润生. 基于注意力机制的堆叠LSTM短时船舶交通流预测模型[J]. 大连海事大学学报, 2024, 50(1): 57-65.
LIAN Q Y, SUN W, LI R S. Stacked LSTMs short-term ship traffic flow prediction model based on attention mechanism[J]. Journal of Dalian Maritime University, 2024, 50(1): 57-65.
[8] 孙杰, 吕文祺, 祁新杰. 船岸连接与船船连接复合通信系统设计[J]. 船舶工程, 2024, 46(S1): 399-403.
SUN J, LYU W Q, QI X J. Design of composite communication system for ship-shore link and ship-ship link[J]. Ship Engineering, 2024, 46(S1): 399-403.
