manage 4D historical AIS data by Space Filling Curve

Abstract

This MSc thesis aims to research how to efficiently manage 4D AIS data (Longitude, Latitude, Time, and MMSI (the ID of the vessel)) to do the fast query by using the Space Filling curve in PostgreSQL. The AIS is the Automatic Identification System which is born because the frequent occurrence of maritime accidents has caused casualties and economic losses. AIS is intended to assist a vessel’s watchstanding officers and allow maritime authorities to track and monitor the vessels’ movement. For now, the AIS has been used in various kinds of fields because the AIS data is really important and useful. The AIS data contains lots of useful information such as the dynamic information (including ship location, speed, heading, and so on), the static information (including ship name, ship type...), and some other types of data. Because of the useful information that AIS includes, there a great many useful applications based on the AIS data. For example, the AIS data is used for detecting the vessels’ anomalies motions or tracking the vessels. While the studies mainly focus on the applications of the AIS data, The efficient management of the AIS data is neglected. Hence, I am going to study how to efficiently manage the multidimensional AIS data. Space Filling Curve (SFC) will be used to manage the multidimensional AIS data. The SFC is a great method for indexing the multidimensional data. The SFC can map data in multidimensional space to 1D space. There are lots of kinds of SFC, such as Morton curve, Hilbert curve, Gray curve, and so on. And the Morton curve and Hilbert curve are used in this thesis because of the property of the locality form the nD space is preserved in the location on the curve [Dai and Su, 2003] and both are so-called quadrant recursive curves [Meijers and van Oosterom, 2018] which is the very significant property. In my research, I proposed two kinds of methods to manage 4D AIS data. One is the 4D integrated approach that the 4D AIS data is encoded to SFC together. The other is the 3D integrated approach, only 3D AIS (Longitude, Latitude, and Time) data is encoded. To test the two approaches, bounding box query (to find the vessels in a given space and time range) and trajectory query (to find the position information of a specific vessel in a give time range) will be implemented in the database. To verify the usability and superiority of my approach, the benchmark is set. The comparison between the two approaches I proposed will be done. And results prove that the SFC approach I used to manage the 4D AIS data is great after comparing it with the benchmark I put forward.