Font Size: a A A

Research On Ship Traffic Flow Prediction Based On EMD-LSTM

Posted on:2023-11-03Degree:MasterType:Thesis
Country:ChinaCandidate:Z JiFull Text:PDF
GTID:2532307040479834Subject:Traffic Information Engineering & Control
Abstract/Summary:PDF Full Text Request
The size of ship traffic flow has always been considered to reflect the scale and busy degree of ship traffic in a certain water area.From the perspective of ensuring maritime traffic safety,ship traffic flow can also reflect the congestion and danger degree of ship traffic in this water area.Therefore,the prediction of ship traffic flow can provide a certain theoretical basis for channel navigation safety assessment or ship navigation management,it can also provide some reference suggestions for the channel supervision department and port and shipping department to take traffic control measures to improve the efficiency of ship navigation.Therefore,the research on ship traffic flow prediction is of certain significance.Ship traffic flow prediction is to establish the corresponding prediction model based on the historical ship traffic flow in a certain water area,so as to predict the size of ship traffic flow in a certain period of time in the future,and its prediction accuracy is the key to reflect the quality of the prediction model.Firstly,this thesis sorts out the characteristics of non-linear and non-stationary ship traffic flow from the research status of ship traffic flow.Secondly,according to the characteristics of non-linear and non-stationary ship traffic flow,a ship traffic flow prediction model combining EMD and LSTM network is proposed,the actual ship traffic flow data of the main channel of Qingdao port is used for example verification.Compared with the prediction results of ARIMA model and LSTM model,it shows that the prediction accuracy of EMD-LSTM model is higher.Finally,the EMD-LSTM model is used to predict the ship traffic flow in the main channel of Qingdao port in the next24 hours.At the same time,the K-means clustering algorithm is used to divide the prediction results into "low peak","flat peak" and "peak" periods,so as to achieve the purpose of identifying the traffic status of the future channel.Through the prediction and cluster analysis of ship traffic flow in the main channel of Qingdao port,it not only provides a certain theoretical basis for the research on ship navigation safety assessment in this water area,but also provides some reference suggestions for relevant maritime management departments to take traffic control measures to improve the efficiency of ships entering and leaving the port.
Keywords/Search Tags:Ship Traffic Flow Prediction, Empirical Mode Decomposition, Long Short Term Memory, Traffic State Recognition
PDF Full Text Request
Related items