摘要
[目的]为提高船舶航迹预测精度和计算效率,解决传统方法容易出现的梯度爆炸、计算时间长等问题,提出基于改进的贝叶斯优化算法与时间卷积神经网络的航迹预测模型。[方法]首先,通过引入时间模式注意力机制,提取各输入特征的权重,保证航迹历史数据的时序性,同时引入可逆残差网络,减少时间卷积神经网络模型训练过程中占用的内存;然后,再采用贝叶斯优化算法对时间卷积网络中的超参数(内核大小K、膨胀系数d)进行寻优;最后,采用五折交叉验证方法对模型进行验证,获得最优模型后进行航迹预测。[结果]采用AIS采集的航迹数据验证,结果表明,在弱耦合、中耦合和强耦合航迹预测中,均方根误差分别平均提高5.5×10^(-5),3.5×10^(-4)和6×10^(-4)。[结论]所提出网络对复杂航迹具有良好的适应性,其预测精度均优于传统模型及LSTM模型,在耦合较强的航迹中仍能保持较高的预测精度。
[Objective]As the traditional ship trajectory prediction method is prone to gradient explosion and long calculation time,this paper seeks to improve its accuracy and calculation efficiency by proposing a ship trajectory prediction model based on an improved Bayesian optimization algorithm (IBOA) and temporal convolution network (TCN).[Method]A temporal pattern attention (TPA) mechanism is introduced to extract the weights of each input feature and ensure the timing of the historical flight track data.At the same time,a reversible residual network (RevNet) is introduced to reduce the memory occupied by TCN model training.The IBOA is then used to find the optimality of the hyperparameters in the TCN (size of kernel K,expansion coefficient d).The model is finally validated using a five-fold cross-validation method,and trajectory prediction is carried out after obtaining the optimal model.[Result]The trajectory data is collected by automatic identification system (AIS) and verified.The root mean square error (RMSE) is found to be increased by5.5×10~(-5),3.5×10~(-4) and 6×10~(-4) in weak coupling,medium coupling and strong coupling track prediction respectively.[Conclusion]The proposed network has good adaptability to complex trajectories and higher accuracy than the traditional model and long short-term memory (LSTM) model,while maintaining high prediction accuracy for trajectories with strong coupling.
作者
李金源
朱发新
滕宪斌
毕齐林
LI Jinyuan;ZHU Faxin;TENG Xianbin;BI Qilin(School of Ship and Marine,Zhejiang Ocean University,Zhoushan 316022,China;College of Marine Engineering,Guangzhou Maritime College,Guangzhou 510725,China)
出处
《中国舰船研究》
CSCD
北大核心
2024年第6期303-316,共14页
Chinese Journal of Ship Research
基金
浙江省大学生科技创新活动计划资助项目(2023R411044)。
关键词
导航
神经网络
贝叶斯优化算法
时间卷积网络
时间模式注意力机制模块
可逆残差网络
AIS数据
navigation
neural networks
Bayesian optimization algorithm
temporal convolutional net-works
temporal pattern attention mechanism module
reversible residual networks
AIS data
作者简介
李金源,男,1997年生,硕士生。研究方向:船舶智能控制,数据挖掘。E-mail:1297437371@qq.com;通信作者:朱发新,男,1982年生,硕士,副教授。研究方向:船舶安全与污染控制。E-mail:zhufax7758@163.com。
