Research On Mining Algorithms Of Ship Berthing Behavior Characteristics Based On AIS Data

The key to berthing alongside,for ocean-going vessel,is to reasonably control three key elements of speed,position,angle between true heading and wharf axis,to ensure the ship entering berth safely,successfully and efficiently.Mining and discovering those three key elements' inherent law and correlation is the precondition for berthing safely.The domestic pilot and captain previously have carried out a lot of researches on berthing.They mainly use their own experiences to present qualitative analysis around piloting and handling techniques,ship control characteristics,based on specific example of ship berthing.To develop in-depth study on universal law of three elements while berthing mentioned above,this paper is going to mine data of berthing ship from a vast ocean of AIS data,then carrying out quantitative research,so that information that reflect latent principles of ships movement characteristics can be gathered.This paper is based on AIS data,combined with marine traffic engineering,database and statistics knowledge.Data mining methods and techniques are used to develop research on ship berthing behavior characteristics.The main works are as follow.(1)To gather,decode and store ship AIS into database by using the Integrated Vessel Information Service System,then performing efficient data cleansing operations in the AIS database,laying a solid foundation of data for subsequent research of ship berthing behavior characteristics.(2)Propose broad and narrow concept of ship berthing behavior based on theory of marine traffic engineering.Then putting forward characteristics of ship berthing behavior by synthesizing the concept of broad theory of ship berthing behavior and three elements of berthing handling,thereby modelling OD track distance of berthing ship.(3)Based on the characteristics of track points in the model of OD track distance of berthing ship,designing algorithms on SQL Server,which includes algorithm of querying AIS data in a specific area,algorithm of OD track distance of berthing ship,mining algorithm of rate of SOG and true heading.Meanwhile,the mid-latitude sailing,using for calculating the distance between each adjacent track point,is embedded into the algorithm of OD track distance of berthing ship.And then dimensionlessly standardizing those calculated distances,highlighting the universality of ship berthing behavior characteristics.(4)Select Xiamen Haicang and Songyu,Tianjin Beijiang and Dongjiang as object port,and select containers whose length is longer than 200 meters as object ship.Those berthing containers are divided into four kinds according to the lenth range [200,250],(250,300],(300,350] and(350,400],namely small size,medium size,large size and ultra-large size.And analyze the berthing behavior characteristics focusing on medium size,large size and ultra-large size containers,obtaining recommended range of relevant dynamic parameters while ship berthing.The results show that mining algorithms of ship berthing behavior characteristics can effectively extracting relevant data,and displaying inherent law of behavior characteristics of different types of ships while berthing as well as the correlations between ship berthing behavior characteristics.That could lay the foundation for further research on other dynamic parameters in course of berthing handling and as well to provide evidence for the competent authority to discover berthing handling habits of diffenrent types of ships.In addition,it could also provide the possibility of making matching plans of dynamic parameters in the course of berthing handling,suitable for different types of ships,for pilot or self-pilot personnel.