船舶轨迹预测是智能航运系统的核心技术之一。现有船舶轨迹预测方法较少考虑目标间运动的相互影响，导致预测准确率低且计算量大。为解决上述问题，提出一种基于组稀疏长短期记忆（Sparse Group Long short-term memory Network，SGLNet）模型的船舶轨迹预测方法。利用编码层对输入的目标船舶运动轨迹数据进行编码，通过长短期记忆网络捕捉每个目标的运动特征；基于笛卡尔积构建水面目标周身网格空间，建立感知社交池化层，共享空间近端目标的隐藏状态；设计基于稀疏表示的掩码模型，对SGLNet网络参数量进行压缩。实验结果表明，相比其他序列预测网络模型，船舶轨迹预测精度提高了55.8%，模型参数量降低了12.77%。该方法满足了水面态势感知中对于船舶轨迹的需求，为构建智能航运系统提供了新的技术路径。
Ship trajectory prediction constitutes one of the core technologies in intelligent shipping systems. Nevertheless, the existing methods for ship trajectory prediction seldom consider the mutual influence of movements among targets, leading to low prediction accuracy and considerable computational volume. To address this issue, we propose a novel trajectory prediction method based on the Sparse Group Long Short-Term Memory Network (SGLNet) model. In this approach, the encoding layer processes input data from of ship movement trajectories, while the LSTM network extracts the motion features of individual targets. A grid space surrounding the ship is constructed using the Cartesian product, and the hidden states of spatially proximal targets are shared to create perceptual social pooling layer. Furthermore, a sparse representation-based mask model is introduced to compress network weights. The proposed method effectively captures interdependent motion dynamics. Experimental results demonstrate that it achieves 55.8% higher accuracy compared to existing models. It offers a novel technical approach for the establishment of an intelligent shipping system.
2026,48(3): 138-144 收稿日期：2025-4-17
DOI：10.3404/j.issn.1672-7649.2026.03.022
分类号：U665.2
基金项目：国家自然科学基金青年科学基金（C类）（52301402）；上海交通大学深蓝计划资助项目（SL2022MS002）；广东省基础与应用基础研究基金资助项目（2022A1515110574）
作者简介：王中瑞(2001-),男,博士研究生,研究方向为无人艇态势感知
参考文献：
[1] 许洋. 船舶轨迹预测与路径规划技术研究[D]. 杭州: 杭州电子科技大学, 2022.
[2] 胡启慧. 智能汽车周边多目标行为识别与轨迹预测[D]. 镇江: 江苏大学, 2023.
[3] HUANHUAN L, HANG J, ZAILI Y. Ship trajectory prediction based on machine learning and deep learning: A systematic review and methods analysis[J]. Engineering Applications of Artificial Intelligence, 2023, 126(C): 418-423.
[4] JASKOLSKI, KRZYSZTOF. Automatic identification system (AIS) dynamic data estimation based on discrete kalman filter (KF) algorithm[J]. Scientific Journals of the Naval Academy, 2017, 211(4): 71-87.
[5] FOSSEN S, FOSSEN T I. Exogenous kalman filter (XKF) for visualization and motion prediction of ships using live automatic identification systems (AIS) data[J]. Modeling, Identification and Control, 2018, 4(39): 233-244.
[6] PERERA L P, OLIVEIRA P, GUEDESSOARES C. Maritime traffic monitoring based on vessel detection, tracking, state estimation, and trajectory prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(3): 1188-1200.
[7] QIAO S J, HAN N, ZHU X, et al. A dynamic trajectory prediction algorithm based on kalman filter[J]. Acta Electronica Sinica, 2018, 46(2): 418-423.
[8] TAMER B. A new approach to linear filtering and prediction problems[C]//Control Theory: Twenty-Five Seminal Papers (1932-1981). IEEE, 2001.
[9] MAO S, TU E, ZHANG G, et al. An automatic identification system (AIS) database for maritime trajectory prediction and data mining[C]//Proceedings of ELM-2016. Springer, 2018.
[10] 徐婷婷, 柳晓鸣, 杨鑫. 基于BP神经网络的船舶航迹实时预测[J]. 大连海事大学学报, 2012, 38(1): 9-11.
[11] SUO Y F, CHEN W K, CLARAMUNT C, et al. A ship trajectory prediction framework based on a recurrent neural network[J]. Sensors, 2020, 20(18): 5133.
[12] JASEENA K U, KOVOOR B C. Decomposition-based hybrid wind speed forecasting model using deep bidirectional LSTM networks[J]. Energy Conversion and Management, 2021, 234: 113944.
[13] HAO X, DU Q, REYNOLDS M. SS-LSTM: A hierarchical LSTM model for pedestrian trajectory prediction[C]//2018 IEEE Winter Conference on Applications of Computer Vision. IEEE, 2018.
[14] ZHAO J S, YAN Z W, ZHOU Z Z, et al. A ship trajectory prediction method based on gat and LSTM[J]. Ocean Engineering, 2023, 289(1): 116159.
[15] BOYD S, PARIKH N, CHU E, et al . Distributed optimization and statistical learning via the alternating direction method of multipliers[C]//IEEE, 2011.
[16] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
[17] NOCEDAL J, WRIGHT S J. Numerical optimization[M]. Springer, 2006: 1–684.
[18] 黄时星. 野外环境下多运动目标检测与轨迹预测方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2023.
[19] 牛鑫, 吕现伟, 余辰. 边缘智能: 现状与挑战[J]. 武汉大学学报(理学版), 2023, 69(2): 270-282.
[20] QIAO S, TANG C, JIN H, et al. PutMode: prediction of uncertain trajectories in moving objects databases[J]. Applied Intelligence, 2010, 33: 370-386.
[21] GAO R, SÄRKKÄ S, CLAVERIA-VEGA R, et al. Autonomous tracking and state estimation with generalized group lasso[J]. IEEE Transactions on Cybernetics, 2022, 52(11): 12056-12070.
