基于特征注意力机制的RNN-Bi-LSTM船舶轨迹预测

Ship Trajectory Prediction of RNN-Bi-LSTM Based on Characteristic Attention Mechanism

【目的】为更准确预测船舶轨迹,基于RNN、Bi-LSTM和注意力机制,研究一种结合特征注意力机制的RNN-Bi-LSTM的船舶轨迹预测模型。【方法】基于AIS数据构建基于循环神经网络(RNN)与双向长短时记忆网络(Bi-LSTM)的混合神经网络模型,并在混合模型中加入特征注意力机制对数据特征进行权重分配,提升模型对船舶轨迹预测精度。【结果】使用实际运行的船舶AIS数据,对模型的有效性和实用性进行验证,测试集均方误差为2.751×10^(-5)、均方根误差为5.245×10^(-3),在连续弯道预测中的均方误差为4.359×10^(-6)、均方根误差为2.088×10^(-3)。【结论】结合特征注意力机制的RNN-Bi-LSTM相较于传统的预测神经网络,船舶轨迹预测精度更高,尤其在弯道预测中也表现出较好的符合度。

【Objective】In order to predict ship trajectory more accurately,based on RNN,Bi-LSTM and attention mechanism,a ship trajectory prediction model of RNN-Bi-LSTM combined with characteristic attention mechanism is studied.【Method】Based on AIS data,a hybrid neural network model based on recurrent neural network(RNN)and bidirectional long/short-term memory(Bi-LSTM)is constructed,and the characteristic attention mechanism is added to the hybrid model to assign weights to the data features,which improves the trajectory prediction of the model.【Result】The effectiveness and practicability of the model AR-Bi-LSTM were verified by using the actual ship AIS data;the mean squared error(MSE)of the test set was 2.751×10^(-5),the root mean square error(RMSE)of the test set was 5.245×10^(-3);while the MSE was 4.359×10^(-6),and the RMSE was 2.088×10^(-3)in the continuous curve prediction.【Conclusion】Experimental results show that the RNN-Bi-LSTM combined with the Characteristic Attention Mechanism has higher accuracy than the traditional prediction neural network,especially in the curve prediction.