Managing Historic Automatic Identification System data by using a proper Database Management System structure.
Presentation: The 4th of November at 15.45 in room BK-IZ U
This MSc thesis aims to develop a management solution for historical Automatic Identification Structure (AIS) data by finding a suitable database and data(base) structure for an effective extraction of historical AIS data to support Spatial-Temporal analyses. The work-flow the thesis proposes focusses on the selection of a database, the organization of the data inside this database and the effectiveness of the extraction of information that can be used in Spatial-Temporal analyses. AIS data, vessel movement data, has a complex structure consisting of 27 different kind of encoded messages. These messages hold their own unique and similar data and are related to each other by their vessel-ID number (MMSI). The update rate of the messages, the data, is dynamic, different for each kind of message. Storing and structuring historic AIS data should be able to handle the data features.
Based on the comparison between the AIS data features, the requirements that the database has to meet which are set by Rijkswaterstaat and related literature and documentation is MongoDB a suitable database to manage AIS data. The requirements that are set by Rijkswaterstaat are based on three use cases; location, trajectory and bounding box. Extraction of the information, that can answer the use cases, from the historic AIS data that is stored within the database, should be effective.
To ensure this, two data(base) organizations are developed, one storing AIS data in a way that it will support the Spatial-Temporal analyses of the specified use cases and one storing AIS data in a way that it will support all possible spatial-temporal use cases. The first will store only the four decoded attributes that are necessary to answer the use cases together with the original encoded AIS message. The second will store all decoded attributes an AIS message holds.
To enhance the performance of MongoDB holding AIS data in one of the two data(base) organization, four indexes on individual attributes and one 4D Morton code index are developed. A 4D Morton code index based on latitude, longitude, MMSI and date-time is developed to enhance the effectiveness of the extraction of the information from the database that will answer the three use cases.
A comparison between the two data(base) organizations and the five indexes by measuring their effectiveness while executing three, on the by Rijkswaterstaat specified use cases based, queries, concludes on a suitable management solution for historical AIS data which will support Spatial- Temporal analyses.
Main mentor: C.W. Quak
Second mentor: Prof. dr. ir. P.J.M. van Oosterom
Co-Reader: J.W. Konings