AIS-Based Vessel Traffic Flow Prediction Using Combined EMD-LSTM Method

The accurate prediction of vessel traffic flow is an essential problem of marine intelligent transportation systems. Existing approaches for predicting vessel traffic flow focus primarily on the trend of historical traffic flow, ignoring the influence of randomness in traffic flow on the prediction of vessel traffic flow. To achieve high precision vessel traffic flow prediction, this study introduced a vessel traffic flow prediction approach based on empirical mode decomposition (EMD) and long-short term memory network (LSTM). Specifically, this paper firstly extracts the traffic flow of vessel traffic by using automatic identification system (AIS); Secondly, in an attempt to reduce the influence of randomness in traffic flow prediction approach, in this study, the vessel traffic flow is decomposed using the EMD algorithm and the Intrinsic mode functions (IMF) of the change in vessel traffic flow are extracted; Then, the LSTM approach is applied to predict multiple IMFs of vessel traffic flow, and the results are superimposed to obtain accurate vessel traffic flow results; Finally, in this paper, we conduct experiments on a huge quantity of AIS data, and the experimental results show the superior performance of the proposed method in vessel traffic flow prediction.