Research On Identification Of Ship Following Behavior In Channel

With the continuous development of maritime informatization,ship trajectory research based on the Automatic Identification System(AIS)of the ship has achieved remarkable results,but the research of ship following is rarely concerned.As an important part of traffic flow theory,ship following theory has important theoretical and application value.In this thesis,in order to explore the following behavior of ships in the channel,the experiment was carried out based on the AIS data of the ship in the Changshan Channel and the car following theory.The similarity measurement effect of the Longest Common Sub-Sequence(LCSS),Edit Distance on Real Sequence(EDR)and Dynamic Time Warping(DTW)is compared.The clustering effects of K-means clustering,spectral clustering and Density-Based Spatial Clustering of Applications with Noise(DBSCAN)are compared and analyzed A method for identifying the following ships in the channel based on improved DTW and DBSACAN is proposed.The results are displayed in the form of visual ship speed time series and ship trajectory space-time chart.The results show that this method can effectively identify the following ships in the channel.It provides support for ship behavior research,ship traffic flow prediction and ship traffic flow stability analysis based on ship following theory.The results of this paper are as follows:(1)Taking the car following theory as a prototype and combining the characteristics of ship traffic,the concept of ship following behavior is proposed.(2)A method for identifying the car following behavior in the channel using the similarity of the ship-to-ground speed change curve as an index is proposed.