This paper presents a method to detect Point Merge patterns from track data. Point Merge is a technique for sequencing arrival flows developed by the EUROCONTROL Experimental Centre, which is now in operation in several places around the world. The motivation for this work is to keep track of its development and the way it is used, to maintain a global picture and a repository of best practices.The method, developed iteratively, exploits geometrical information to detect Point Merge signature patterns. This procedure has been tested on the European data set (2230 airports, one month) obtaining good detection performance (8 correct detection out of 10 known implementations) in regular operation modes and false alarm or miss for low traffic conditions. Furthermore, we have analyzed a worldwide data set from FlightRadar24 (900 busiest airports, one week) that allowed to identify 3 new Point Merge implementations outside Europe, and confirmed 8 already known.Future work will focus on data-driven techniques and the use of machine learning to obtain better discriminating features and improve the pattern detection scheme. Moreover, the proposed procedure will be run over the different airports in order to maintain the list of Point Merge implementations and better understand the similarities and differences among the different usages in different locations.
Detecting Point Merge Patterns From Track Data
2020-09-01
4804198 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
DETECTING POINT MERGE PATTERNS USING COMPUTER VISION
TIBKAT | 2021
|Detecting Point Merge patterns using computer vision
AIAA | 2021
|Data-adapted Parallel Merge Sort
Deutsches Zentrum für Luft- und Raumfahrt (DLR) | 2019
|