Modeling And Spatio-temporal Characteristics Research Of Vessel Conflict In Confluence Waters

In recent years,the number of vessels sailing along the coast of China has increased day by day,and has been increasing toward large-scale and specialized.The increase of the density and flow of vessels in coastal waters has caused frequent conflicts between sailing vessels.There are some deficiencies in the analysis of spatio-temporal characteristics of conflicts in existing vessel conflict studies.In order to excavate the spatio-temporal distribution and temporal-spatial correlation characteristics of ship conflicts in confluence waters more comprehensively and deeply.Based on the theory of spatial autocorrelation,this paper proposes a temporal and spatial statistical model that can simultaneously analyze and represent ship conflict characteristics from both time and space dimensions.In view of the spatio-temporal autocorrelation of ship conflict data,a spatial regression model is constructed to eliminate the regression errors caused by spatio-temporal characteristics and improve the fitting accuracy of the model.The main research of this paper include:(1)The mechanism of ship conflict in confluence waters is analyzed and the mathematical model of ship conflict is constructed.In this paper,the ship conflict between two ships in the confluence waters along the coast is taken as the research object.Firstly,the basic knowledge of ship conflict is summarized.Secondly,the mechanism of ship conflict is analyzed.Then,ship conflict is divided from different angles.Finally,the mathematical model of ship conflict is constructed.Taking the typical intersection waters near the Xiazhimen Channel as an example,the automatic identification and extraction of ship conflict data in AIS data are realized by Python.(2)Extending the theory of spatial autocorrelation in time dimension,and the spatio-temporal analysis model of ship conflict data in confluent waters is constructed.The main part of exploratory analysis includes the analysis of the influence of spatial grid scale,time delay scale and order of spatial weight matrix on spatio-temporal distribution pattern of conflict data,as well as the identification of spatio-temporal hot spots in ship conflicts.On this basis,the spatio-temporal correlation of conflict data is studied by using appropriate spatio-temporal scale and spatial weight matrix.Experiments show that the spatiotemporal autocorrelation of ship collisions decreases with the increase of the distance between spatial grids and the time delay and ship conflict data have obvious spatial heterogeneity and temporal dynamics.(3)Regression analysis of ship conflict data shows that the residual has spatial autocorrelation property.In order to eliminate the influence caused by the spatial autocorrelation of model residuals,two spatial autoregressive models,SEM and SLM,were established.The experimental results show that the regression coefficient of ship length is negative,that is,small ships are more likely to have ship conflicts than large ships;the regression coefficient of the number of cargo ships in the spatial grid is the largest,indicating that the increase of the number of cargo ships in the water will increase the probability of ship conflicts.The automatic identification system of ship conflict implemented in this paper provides a data base for the study of spatio-temporal characteristics of ship conflict.The spatio-temporal correlation analysis of conflict data provides a new method for identifying risk areas in waters.At the same time,the spatial regression analysis of ship conflict data can effectively analyze the impact of various ship traffic parameters on ship conflict.