University of Twente Student Theses


Back End Support for Trajectory Management Using Open Source Software

Iswari, Lizda (2010) Back End Support for Trajectory Management Using Open Source Software.

[img] PDF
Abstract:Trajectory is defined as a movement of an object in space and time. It consists of a set of spatial and temporal data. Exploration of this data set can be used to express the movement behaviour of moving objects that is previously unknown. Trajectory data exploration can be conducted based on their geometric and semantic properties. The geometric properties related to the characteristics of the trajectory are travelled time, travelled distance, average speed and direction of movement. The sematic properties can be obtained from the requirements and needs of users of the trajectory data. This research proposes several alternatives on how to manage and explore trajectory data in a database system. Research has been conducted in two stages. The first stage deals with designing the data models. There are two models involved, a general trajectory model and an application-specific data model. These data models are designed as UML class diagrams that specify the required attributes and methods to manage trajectories in a database system. The second stage is associated with the provision of back end support to extract the geometric and the semantic properties of trajectory. The iceberg movements in Antarctica have been chosen as a case study. Based on the analysis of user requirements, there are some necessities deal with: managing data errors in the data set, extracting geometric properties of the iceberg, detecting iceberg events and analyzing behaviour movement of icebergs. To meet these requirements, four type of database functions were implemented in the PostgreSQL database system. The first functions deal with data pre-processing, i.e. to clean data set from inconsistencies, empty values, duplicates and outliers. The second functions deal with extracting the characteristic of an iceberg. The third functions deal with classification of event during the lifespan of an iceberg and functions to detect the calving occurrences. The last functions deal with trajectory data mining to find similar patterns of iceberg movement based on some criteria that are determined by the users. All these functions still consider trajectory as an individual entity. Further research is needed to provide back end support that also involve neighbourhood information and relationship to other trajectories that enable data exploration based on spatial and temporal proximity. Keywords back end support, iceberg event, data pre-processing, similar pattern, trajectory data mining
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page